Guild icon
OpenRouter
💬 chat / models / Momentum
Avatar
movementlabsAI 11/3/2025 2:49 PM
Momentum might be coming soon to everyone via OpenRouter.😀
👀 9
Avatar
Is the model available somewhere, or?
Avatar
where did you got that data neither gpt-5 and sonnet 4.5 runs at 120? and gemini models runs on google tpus not on gpus
2:55 PM
doesn't matter because that model is crazy fast anyway (edited)
3:00 PM
idk how you guys made it but if that partnership goes forward i would probably use that model for a lot of things
Avatar
Avatar
Biel
where did you got that data neither gpt-5 and sonnet 4.5 runs at 120? and gemini models runs on google tpus not on gpus
movementlabsAI 11/3/2025 3:02 PM
Thank you for pointing this out, this will be fixed shortly, yes google does indeed use TPUs, we just launched today and are going over website ui sections. Good catch 👍
Avatar
Avatar
movementlabsAI
Momentum might be coming soon to everyone via OpenRouter.😀
looks interesting, the benchmark numbers are especially high is there any free credits or api?
Avatar
Avatar
toriset
looks interesting, the benchmark numbers are especially high is there any free credits or api?
there is an api on openai format
3:05 PM
but idk about free api credits
3:05 PM
but you can use their chat for 50 msg/day
Avatar
hmm ok
Avatar
the model is not too bad on frontend that is plain html
Avatar
Avatar
Biel
the model is not too bad on frontend that is plain html
movementlabsAI 11/3/2025 3:11 PM
Glad you to see you diving right into coding, we are focused on speed on this launch, next launch in the coming days we are going to update our model with our new checkpoint and it one of the best coding models we have ever seen.
Avatar
Avatar
Biel
the model is not too bad on frontend that is plain html
pretty nice at webdev yeah
Avatar
Avatar
toriset
pretty nice at webdev yeah
here is the file that it generated
24.71 KB
Avatar
i asked it to maek a website of it self and it looks pretty clean imo though im not sure whether those wallets are real
15.02 KB
3:13 PM
👀 1
Avatar
it likes light theme
3:14 PM
a lot
3:14 PM
idk why
3:15 PM
@movementlabsAI that model have reasoning?
3:16 PM
or is it hidden? (edited)
Avatar
Avatar
Biel
@movementlabsAI that model have reasoning?
movementlabsAI 11/3/2025 3:16 PM
This is not a reasoning model, however the updated one will be reasoning, however it will decide when to reason and we will also allow users to disable reasoning.
Avatar
Avatar
movementlabsAI
This is not a reasoning model, however the updated one will be reasoning, however it will decide when to reason and we will also allow users to disable reasoning.
like the gpt-5 router?
3:17 PM
you guys train it on the mpu too?
3:18 PM
like the tps speed affects training?
Avatar
Avatar
Biel
you guys train it on the mpu too?
movementlabsAI 11/3/2025 3:18 PM
Great question MPU doesn't support training at the moment, MPU is a chip that loads model into the memory.
Avatar
Avatar
movementlabsAI
Great question MPU doesn't support training at the moment, MPU is a chip that loads model into the memory.
the mpu can just run the model not train it
3:18 PM
right?
Avatar
Avatar
Biel
the mpu can just run the model not train it
movementlabsAI 11/3/2025 3:19 PM
Exactly! you nailed it.
Avatar
theres a small typo on your guys' website on the contact page it says 2$/M In while on the pricing page its 3$/M in
👍 1
Avatar
you should reach to artificial analysis offer then credits so they can benchmark the model https://artificialanalysis.ai/models
Comparison and analysis of AI models across key performance metrics including quality, price, output speed, latency, context window & others.
👍 3
Avatar
Avatar
toriset
theres a small typo on your guys' website on the contact page it says 2$/M In while on the pricing page its 3$/M in
movementlabsAI 11/3/2025 3:24 PM
Movement Labs AI
Avatar
Avatar
movementlabsAI
We fixed the stats https://www.movementlabs.ai/mpu
it should be lower not higher
3:25 PM
gpt-5 from openai gets a solid 40
3:25 PM
sonnet 4.5 smt like 60
Avatar
also grok 4 runs at like 30 t/s unless its grok 4 fast which is like 150
Avatar
Avatar
toriset
also grok 4 runs at like 30 t/s unless its grok 4 fast which is like 150
yea
3:26 PM
gemini pro is ~100
3:26 PM
the flash get ~300
Avatar
Avatar
toriset
also grok 4 runs at like 30 t/s unless its grok 4 fast which is like 150
i don't really play with x.ai models they are expensive and dumb at the same time the 4 fast is fast but is dumber then haiku 4.5
Avatar
Avatar
Biel
i don't really play with x.ai models they are expensive and dumb at the same time the 4 fast is fast but is dumber then haiku 4.5
grok 4 fast is very cheap and fast, but depends on the use case if its good
Avatar
Avatar
toriset
theres a small typo on your guys' website on the contact page it says 2$/M In while on the pricing page its 3$/M in
movementlabsAI 11/3/2025 3:28 PM
Fixed 🙂
Avatar
Avatar
toriset
grok 4 fast is very cheap and fast, but depends on the use case if its good
yea like simple things it is good but you give it more things to do it just gets lost
3:29 PM
and grok models hallucinate a lot of tool calls
3:29 PM
haiku and gpt-5-codex too
Avatar
Avatar
Biel
and grok models hallucinate a lot of tool calls
noticed this too, they think they did the tool call during reasoning but it didnt actually trigger (maybe inference problem rather than the model)
Avatar
Avatar
toriset
noticed this too, they think they did the tool call during reasoning but it didnt actually trigger (maybe inference problem rather than the model)
yea idk
3:32 PM
the bottleneck on that model will be the tools lol
3:32 PM
and bandwidth
Avatar
movementlabsAI 11/3/2025 3:36 PM
Momentum follows tools.
Avatar
yea the tool is the bottleneck
3:38 PM
you can see the delay
Avatar
movementlabsAI 11/3/2025 3:38 PM
That's from cursor, on roo code and cline it's much faster
Avatar
that model runs only on 1 mpu or it needs more?
3:38 PM
cuz 1 have 64gb
Avatar
movementlabsAI 11/3/2025 3:39 PM
Currently we are running it on a few, once we get some traction we will scale it up to 100-200 MPUs
Avatar
Avatar
movementlabsAI
Currently we are running it on a few, once we get some traction we will scale it up to 100-200 MPUs
that will be crazy
Avatar
Avatar
Biel
that will be crazy
movementlabsAI 11/3/2025 3:40 PM
Game changing for many, we're glad to have you here testing early.
Avatar
can we get some api credits to test the models in something like cline?
Avatar
https://movementlabs.ai/about the founder video is broken
Movement Labs AI
Avatar
Avatar
Biel
https://movementlabs.ai/about the founder video is broken
movementlabsAI 11/3/2025 3:41 PM
yes, sure, please DM me your email i will issue you some free credits.
Avatar
Avatar
movementlabsAI
yes, sure, please DM me your email i will issue you some free credits.
done :)
Avatar
can't send
3:42 PM
discord says that it can't send messages
3:42 PM
sorry
3:42 PM
error on my side
Avatar
movementlabsAI 11/3/2025 3:44 PM
Avatar
What country are you guys based at?
Avatar
Avatar
Biel
What country are you guys based at?
movementlabsAI 11/3/2025 3:47 PM
United Kingdom, just two founders. With a big vision
Avatar
idk if you can share how much parameters the model have, but if you can i just want to know
Avatar
Avatar
Biel
idk if you can share how much parameters the model have, but if you can i just want to know
movementlabsAI 11/3/2025 4:16 PM
I can tell you it's above 700 billion running FP16 (half-precision) - Hope that helps 🙂
Avatar
Avatar
movementlabsAI
Click to see attachment 🖼️
cairo com 2 a & 2 i 11/3/2025 4:17 PM
wow, that's crazy fast
4:17 PM
looking forward to using it in OpenRouter
🤝 1
Avatar
movementlabsAI 11/3/2025 4:20 PM
We welcome constructive feedback to help improve our training checkpoints and model performance.
Avatar
Avatar
movementlabsAI
I can tell you it's above 700 billion running FP16 (half-precision) - Hope that helps 🙂
wow thats huge
Avatar
Avatar
movementlabsAI
We welcome constructive feedback to help improve our training checkpoints and model performance.
If you weren't claiming better performance than GPT-5 we could have considered
Avatar
Avatar
Cocc
If you weren't claiming better performance than GPT-5 we could have considered
it could be worse then gpt-5 but you can iterate much quicker
Avatar
Avatar
movementlabsAI
I can tell you it's above 700 billion running FP16 (half-precision) - Hope that helps 🙂
Oh wow, the speed's insane if so! Congrats on the release I wonder how fast it could run a 30B model, but that's just mad scientist-me speaking LMAO
🩷 1
Avatar
You can't convince me with that speed they didn't have "time" to do a single major benchmark
Avatar
Avatar
Cocc
If you weren't claiming better performance than GPT-5 we could have considered
movementlabsAI 11/3/2025 4:43 PM
Keep in mind there are our internal benchmarks, we invite you to be your own judge.
Avatar
Avatar
MetalDragon
You can't convince me with that speed they didn't have "time" to do a single major benchmark
There are these ones listed, though they're averaged out so that makes it harder to reproduce
4:44 PM
The video here doesn't play for me: https://www.movementlabs.ai/about
Avatar
that model feels like sonnet 4.5 is some ways
Avatar
Avatar
KP
The video here doesn't play for me: https://www.movementlabs.ai/about
movementlabsAI 11/3/2025 4:46 PM
We are going to upload it to youtube and replace that section with the youtube embed.
Avatar
Avatar
Biel
that model feels like sonnet 4.5 is some ways
the way that it uses emojis and talks really feels like claude
Avatar
movementlabsAI 11/3/2025 4:51 PM
Glad you are enjoying it, the speed is a bonus!
Avatar
does it support images?
Avatar
Wow this is actually really cool
4:57 PM
Surprised!
4:58 PM
Is this a fresh pretrained model? Or is it based on some Qwen base or something?
Avatar
Avatar
Bulbasaur
Is this a fresh pretrained model? Or is it based on some Qwen base or something?
i think that it is a fresh one
4:59 PM
no model does frontend in the same style
Avatar
thats very impressive if it is, it takes a hell lot of time to train a 700B param model from scratch, especially if your hardware isnt designed for that
Avatar
>above 700B yeah huh
Avatar
Avatar
toriset
thats very impressive if it is, it takes a hell lot of time to train a 700B param model from scratch, especially if your hardware isnt designed for that
Yeah
5:00 PM
I don’t know of any open weights model above 700B but <1T
Avatar
Avatar
Cocc
If you weren't claiming better performance than GPT-5 we could have considered
movementlabsAI 11/3/2025 5:00 PM
We are about to send a video here that will show some facts. Incoming.
Avatar
Avatar
Bulbasaur
I don’t know of any open weights model above 700B but <1T
deepseek is close at 671 or sum around that
Avatar
Avatar
toriset
deepseek is close at 671 or sum around that
Yea
Avatar
it could be a 1T but pruned and finetuned
Avatar
Avatar
toriset
it could be a 1T but pruned and finetuned
yea but 1t at 2400 tps sounds crazy
Avatar
Avatar
movementlabsAI
We are about to send a video here that will show some facts. Incoming.
How much will mTok cost be?
Avatar
Avatar
Biel
yea but 1t at 2400 tps sounds crazy
eh its probably MoE so it makes it much easier (edited)
Avatar
Avatar
Bulbasaur
How much will mTok cost be?
3$ in 5$ out (edited)
👍 1
Avatar
Avatar
toriset
eh its probably MoE so it makes it much easier (edited)
yea
Avatar
Avatar
Biel
3$ in 5$ out (edited)
isnt it 3$/m in and 5$/m out?
Avatar
Avatar
toriset
isnt it 3$/m in and 5$/m out?
yea sorry
Avatar
Avatar
Cocc
If you weren't claiming better performance than GPT-5 we could have considered
movementlabsAI 11/3/2025 5:02 PM
Avatar
messd up with haiku pricing
Avatar
Avatar
movementlabsAI
prety cool physics
Avatar
Avatar
movementlabsAI
both are decent results, and gpt 5 is technically the reasoning model not the chat version, but this model is good for a non reasoning model
Avatar
Avatar
Biel
messd up with haiku pricing
movementlabsAI 11/3/2025 5:04 PM
We will bring prices down to $1 /m in and $1/m out in the coming weeks.
👀 7
Avatar
i asked the model to create battlefield 6 with html
5:04 PM
29.33 KB
Avatar
@movementlabsAI can you confirm if it’s a fresh pretrained model? Or based on other open weights models?
Avatar
Avatar
movementlabsAI
We will bring prices down to $1 /m in and $1/m out in the coming weeks.
will we get cache support in the visible future?
👆 1
Avatar
Model isn’t hallucinating badly afaict
Avatar
Avatar
movementlabsAI
We will bring prices down to $1 /m in and $1/m out in the coming weeks.
you should get kv cache for the tokens so the input cached would be cheaper
Avatar
Many many open weights models still hallucinate niche knowledge insanely
Avatar
it halucinated a litte with me
5:07 PM
i asked for a ranking on embeding models
5:07 PM
and it said to me that claude had a embeding model
Avatar
it doesn't have any tools to verify
Avatar
@movementlabsAI you should really turn on sign in with Google on your site, should be easy with Clerk that you’re already using :)
Avatar
if you asked claude the same thing without websearch it would probably do the same thing
Avatar
you guys should really make the input on the chat area multi line, its killing me scrolling with 2 lines visible at a time
Avatar
Nvm it’s still hallucinating
5:11 PM
“What is…” Tailwind v4 and React Router v7 (edited)
5:11 PM
Always gets ‘em
5:11 PM
LULW
Avatar
its writting like qwen / deepseek
Avatar
Avatar
toriset
you guys should really make the input on the chat area multi line, its killing me scrolling with 2 lines visible at a time
movementlabsAI 11/3/2025 5:11 PM
On it, noted ! 🙂
Avatar
Avatar
toriset
its writting like qwen / deepseek
that sounds more like claude
Avatar
Avatar
toriset
its writting like qwen / deepseek
movementlabsAI 11/3/2025 5:13 PM
We feel like the writing style is in-between GPT4o and Sonnet.
Avatar
Avatar
Biel
that sounds more like claude
from what ive seen qwen does that a lot with the emojies, often with a bunch of checkmarks
Avatar
Avatar
toriset
from what ive seen qwen does that a lot with the emojies, often with a bunch of checkmarks
claude does the same thing
Avatar
Avatar
Biel
claude does the same thing
hmm
Avatar
let me get a chat with claude
5:14 PM
Avatar
interesting mistake.. almost never seen any model do PI instead of Math.PI in JS
Avatar
its in pt-br
Avatar
Avatar
Biel
Click to see attachment 🖼️
my claude (maybe cuz of my instructions) is way more bland but when i use it in copilot it does kinda type like that
Avatar
Avatar
toriset
my claude (maybe cuz of my instructions) is way more bland but when i use it in copilot it does kinda type like that
yea my claude have some instructions to be direct that is an chat before i put the new instructions (edited)
Avatar
Avatar
Bulbasaur
Click to see attachment 🖼️
that duck is different
Avatar
hes a bit special
Avatar
Avatar
toriset
hes a bit special
a bit?
5:17 PM
a lot
5:18 PM
i need smt hard to ask him
Avatar
make a blackhole in three js
Avatar
Avatar
toriset
make a blackhole in three js
i think taht svete is harder
5:20 PM
Avatar
not the most realiistic thing
Avatar
Avatar
Bulbasaur
@movementlabsAI you should really turn on sign in with Google on your site, should be easy with Clerk that you’re already using :)
movementlabsAI 11/3/2025 5:21 PM
We have enabled google login and signup. 👍
👍 1
Avatar
Avatar
toriset
not the most realiistic thing
the model maybe shine in some backend stuff
Avatar
Avatar
toriset
make a blackhole in three js
movementlabsAI 11/3/2025 5:22 PM
Avatar
a task that i was using to benchmark models was to ask the model to port litellm backend from python to go
5:24 PM
gpt-5 just run the entire night and failed
Avatar
Avatar
Biel
a task that i was using to benchmark models was to ask the model to port litellm backend from python to go
has any model actually succeeded? thats a pretty big task
Avatar
grok code fast 1 was the closest then just looped
5:25 PM
grok code fast 1 was like 80%
5:25 PM
then just looped and didin't made any progress
5:26 PM
claude i didn't tested because i know that will be expensive
5:26 PM
gpt-5 i tested it when it was free on cursor (edited)
Avatar
movementlabsAI 11/3/2025 5:26 PM
We're absolutely loving the amazing community responses we're seeing. The engagement is next level! Quick question for you all should we bump up those daily free message limits? We're curious what you think! 🚀
Avatar
Avatar
movementlabsAI
We're absolutely loving the amazing community responses we're seeing. The engagement is next level! Quick question for you all should we bump up those daily free message limits? We're curious what you think! 🚀
idk i didn't hit any limit (edited)
👍 1
Avatar
Avatar
movementlabsAI
We're absolutely loving the amazing community responses we're seeing. The engagement is next level! Quick question for you all should we bump up those daily free message limits? We're curious what you think! 🚀
can you give my an api key with smt that you store the data thet i put into it so i can try porting the litellm backend to go?
5:30 PM
how long is the model context?
Avatar
200k based on the api docs
Avatar
Avatar
Biel
can you give my an api key with smt that you store the data thet i put into it so i can try porting the litellm backend to go?
movementlabsAI 11/3/2025 5:31 PM
here is an API key for you: ml_5403932ad166f04b67da678bd9c238150f26f1227c2d884e82e0eda6e8164334
🤨 1
5:32 PM
it has $8 worth of credits, try not to burn them in one promt haha 🙂
5:41 PM
the credits just burn so fast
Avatar
Avatar
Biel
Click to see attachment 🖼️
might be a model issue
Avatar
movementlabsAI 11/3/2025 5:44 PM
please check now
5:45 PM
credits did not burn, it was server side config, we updated something on our end maybe cause a glitch during that time.
Avatar
trying again
👍 1
5:48 PM
it's crawling the backend
5:49 PM
for some reason the ttft is high
5:50 PM
@movementlabsAI are you shure that the context is 200k
5:50 PM
cuz the context hit 128k and stalled
5:50 PM
Avatar
movementlabsAI 11/3/2025 5:52 PM
the context is 128k 200k for upcoming model this week, for now just resume task as many times as you like (edited)
5:53 PM
i have added you more credits
5:53 PM
its still on architect
5:54 PM
the model is really good at tool calling
Avatar
movementlabsAI 11/3/2025 5:55 PM
we are egar to see what happens
5:56 PM
you must have a very big codebase
Avatar
yea its the lite llm one
5:56 PM
Python SDK, Proxy Server (LLM Gateway) to call 100+ LLM APIs in OpenAI format - [Bedrock, Azure, OpenAI, VertexAI, Cohere, Anthropic, Sagemaker, HuggingFace, Replicate, Groq] - BerriAI/litellm
5:56 PM
i told him to port the python backend to go
5:57 PM
its 1m+ lines i think
5:57 PM
i will download obs and start a stream on youtube
👀 1
Avatar
movementlabsAI 11/3/2025 5:58 PM
That's pretty big, yes! 🙂
6:01 PM
What do you think of the speed in plan mode?
Avatar
its fast but it is missing some tool calls
Avatar
@movementlabsAI can you check dms?
Avatar
bielpaesxd14 streams live on Twitch! Check out their videos, sign up to chat, and join their community.
6:04 PM
here
6:05 PM
youtube wasn't letting me stream
6:07 PM
that's probably the 5 time its reading the _types
6:11 PM
it really feels like gpt-5
6:11 PM
like
6:11 PM
it reads everything before making one change
Avatar
movementlabsAI 11/3/2025 6:12 PM
i'm watching the stream very intresting
Avatar
would be cool if you could get a 1m context version
6:14 PM
it is just compacting like crazy
Avatar
movementlabsAI 11/3/2025 6:14 PM
If that's what you want we will push for 1m on next version
Avatar
a reasonng one would probably boost more on he benchmark
Avatar
movementlabsAI 11/3/2025 6:15 PM
128k is pushing it for such a large codebase especially to refactor everything 😃
Avatar
movementlabsAI 11/3/2025 6:16 PM
This is a beta release anyway, what we have upcoming is the real deal. To get some feedback going getting the name out there a bit.
👍 1
6:17 PM
to bump the context is it just change a variable or you need to do some training on the model?
Avatar
Avatar
Biel
to bump the context is it just change a variable or you need to do some training on the model?
movementlabsAI 11/3/2025 6:18 PM
Requires some architecture change but fairly straight forward for us.
Avatar
with the speed that it have maybe you can do some wierd stuff with it like make it play some fps game
6:21 PM
the model really condense the context fast
6:22 PM
it's on a loop
Avatar
movementlabsAI 11/3/2025 6:23 PM
i think for a codebase this size it's better to wait for the 1m version - as 128k tokens is not cut for such large codebases but rather smaller guided tasks within the codebase.
Avatar
its on a loop it reads the team endpoints then reads the organization then condenses context
6:24 PM
let me try enabling rag
Avatar
movementlabsAI 11/3/2025 6:24 PM
yes, when it condenses context it looses previous context and that is not good.
Avatar
Avatar
Biel
Click to see attachment 🖼️
cairo com 2 a & 2 i 11/3/2025 6:26 PM
i think claude does that only in its web chat
6:27 PM
qwen models write like that via API with no system prompt, whereas claude models don't
Avatar
Avatar
cairo com 2 a & 2 i
i think claude does that only in its web chat
claude on the chat is the same model from the api
Avatar
cairo com 2 a & 2 i 11/3/2025 6:29 PM
it's the same model, but it has a gigantic system prompt
Avatar
Avatar
cairo com 2 a & 2 i
it's the same model, but it has a gigantic system prompt
that's true
6:30 PM
but i've seen that behavior on my openwebui too
6:32 PM
the embeding is taking a while
Avatar
Avatar
movementlabsAI
yes, when it condenses context it looses previous context and that is not good.
cairo com 2 a & 2 i 11/3/2025 6:33 PM
hey, add your website to your discord bio
Avatar
Avatar
cairo com 2 a & 2 i
hey, add your website to your discord bio
movementlabsAI 11/3/2025 6:33 PM
👍 on it.
Avatar
Avatar
Biel
the embeding is taking a while
movementlabsAI 11/3/2025 6:35 PM
We are super excited for 1m context window, it will really be good for the public, the team and i are discussing this.
Avatar
Avatar
movementlabsAI
We are super excited for 1m context window, it will really be good for the public, the team and i are discussing this.
k i'm embeding the codebase to see if that make the model stop looping
👍 1
6:36 PM
done
Avatar
movementlabsAI 11/3/2025 6:37 PM
Best of luck!
Avatar
it took the search tool too literal
🤖 1
6:40 PM
it needs a bit of stearing it's in the middle of gpt-5 and claude
Avatar
Avatar
Biel
it needs a bit of stearing it's in the middle of gpt-5 and claude
movementlabsAI 11/3/2025 6:44 PM
I think a command is waiting for your approval (edited)
6:45 PM
i went grab smt to eat
6:48 PM
finally some file edits
6:51 PM
it implemented the db
Avatar
movementlabsAI 11/3/2025 6:54 PM
if it had 1m context it would have been a lot better
6:54 PM
overall so far what do you think of the model?
Avatar
for 128k context its great so far
6:56 PM
the go dependency download has failed
Avatar
movementlabsAI 11/3/2025 6:57 PM
i'm following the stream and we are also taking notes.
Avatar
Avatar
movementlabsAI
i'm following the stream and we are also taking notes.
k
6:57 PM
it halucinated a package
6:58 PM
that github.com/BerriAi
6:58 PM
it will probably catch that on the debug phase
Avatar
Avatar
Biel
it will probably catch that on the debug phase
movementlabsAI 11/3/2025 7:00 PM
yes most likely, it will fix itself.
7:00 PM
rag helped
Avatar
Avatar
movementlabsAI
rag helped
a lot
7:02 PM
that is how much the grok code fast 1 used
7:02 PM
that was the beta
7:02 PM
it was free
Avatar
movementlabsAI 11/3/2025 7:03 PM
Did grok code get this far? or did it fail totally? x ai have a lot of backing and VCs funding them 🙂
Avatar
Requests Input Tokens Output Tokens Cache Hits
Avatar
Avatar
movementlabsAI
Did grok code get this far? or did it fail totally? x ai have a lot of backing and VCs funding them 🙂
it made at ~80% then just started a loop
7:03 PM
after that i couldn't make it work properly anymore
Avatar
movementlabsAI 11/3/2025 7:04 PM
and how far are we? with momentum ?
7:04 PM
till now
7:04 PM
maybe 30 40
7:04 PM
but grok had the entire day (edited)
7:05 PM
it fixed itself
7:06 PM
it started doing unit tests
Avatar
Avatar
Biel
it fixed itself
movementlabsAI 11/3/2025 7:07 PM
momentum?
Avatar
Avatar
movementlabsAI
momentum?
yea
7:07 PM
the wrong dependency
7:08 PM
it have 535973 lines to port
7:08 PM
it ported 4313
7:09 PM
grok made ~200-300k
7:09 PM
but it is writing unit tests on it's own
7:09 PM
before saying thats ready
7:10 PM
for a non reasoning model i've never seen one so deterministic
Avatar
Avatar
Biel
for a non reasoning model i've never seen one so deterministic
movementlabsAI 11/3/2025 7:12 PM
That's very nice to hear! This model is coming with reasoning and a 1m context window too
Avatar
Avatar
movementlabsAI
That's very nice to hear! This model is coming with reasoning and a 1m context window too
Both?
Avatar
movementlabsAI 11/3/2025 7:12 PM
yes both baked into this model
Avatar
Avatar
movementlabsAI
yes both baked into this model
the reasoning data will be obfuscated in some way or it will be like claude?
7:13 PM
that they just give you the data
Avatar
movementlabsAI 11/3/2025 7:13 PM
Reasoning tokens will be raw, so users can see them
👍 1
Avatar
Avatar
movementlabsAI
Reasoning tokens will be raw, so users can see them
that's nice
Avatar
movementlabsAI 11/3/2025 7:17 PM
so far your usage is at Input tokens 26,265,035 Output tokens 163,479
🤯 1
7:18 PM
😃
7:18 PM
its very nice to see it in action up againgst such a big task.
Avatar
it started on a loop because kilo code bugged
7:20 PM
then it found a work around with cmd
7:20 PM
it's very token efficient
Avatar
Avatar
Biel
it started on a loop because kilo code bugged
movementlabsAI 11/3/2025 7:21 PM
I guess the IDE matters too, as different IDE's have different tool formats.
Avatar
it's the same cuz kilo build a compat layer for jetbrains
7:22 PM
kilo code is a cline/roo fork
7:23 PM
i like jetbrains ides
7:23 PM
it's starting to fix issues
Avatar
movementlabsAI 11/3/2025 7:23 PM
Web access is something we are working on, it will really help the model too
Avatar
Avatar
movementlabsAI
Web access is something we are working on, it will really help the model too
you just need to train it to recive image input
👍 1
7:24 PM
then you just sends images
7:24 PM
you can make a wrapper of chorme dev tools mcp
7:25 PM
but you still need image input for it to work properly
7:26 PM
i think that kv cache is smt you need to work on to boost the ttft
Avatar
Avatar
Biel
i think that kv cache is smt you need to work on to boost the ttft
cairo com 2 a & 2 i 11/3/2025 7:27 PM
i think that delay is from the IDE itself
Avatar
movementlabsAI 11/3/2025 7:27 PM
this is a server side lag on api calls, since it checks our db for authentication + the IDE.
Avatar
Avatar
cairo com 2 a & 2 i
i think that delay is from the IDE itself
probably the compat layer
Avatar
Avatar
movementlabsAI
this is a server side lag on api calls, since it checks our db for authentication + the IDE.
yea
Avatar
movementlabsAI 11/3/2025 7:28 PM
the api we will make it more efficient to check for api keys
Avatar
a java browser isn't the fastest (edited)
7:29 PM
the ide is just giving up on diffs
7:29 PM
probably because of the speed
7:32 PM
you should sell that mpu to other companies
7:32 PM
you should be able to get some good money with that
Avatar
movementlabsAI 11/3/2025 7:37 PM
The MPU is the heart of our company 🙂
Avatar
Avatar
movementlabsAI
The MPU is the heart of our company 🙂
same with nvidia but they sell them
7:38 PM
there are only 4 companies that i know that does it's own hardware and beats nvidia
7:38 PM
google, groq, cerebras, you
7:38 PM
i think that cerebras sell but idk
Avatar
cairo com 2 a & 2 i 11/3/2025 7:39 PM
lmao perplexity/sonnet 4.5 is EXTREMELY skeptical of your chip
Avatar
Avatar
cairo com 2 a & 2 i
lmao perplexity/sonnet 4.5 is EXTREMELY skeptical of your chip
what they said?
Avatar
cairo com 2 a & 2 i 11/3/2025 7:40 PM
first they've confused Movement with a crypto company
7:41 PM
hey i'm not judging you MovementLabs, i jsut went researching and didn't expect those responses from perplexity
🤝 1
7:41 PM
i don't even know how Cerebras or Groq or even Google get those speeds, i'm not really knowledgeable on this
Avatar
Avatar
cairo com 2 a & 2 i
first they've confused Movement with a crypto company
movementlabsAI 11/3/2025 7:42 PM
I would refrain from asking another AI model about another AI model 😁 it's all very biased
Avatar
cairo com 2 a & 2 i 11/3/2025 7:42 PM
not really, especially with grounding search
Avatar
movementlabsAI 11/3/2025 7:43 PM
how we do it is we scale our chip and load the entire model into memory, that's where the speed comes from
7:43 PM
GPUs, TPU's fetch model shards in and out of memory
Avatar
cairo com 2 a & 2 i 11/3/2025 7:44 PM
i see, the 63.4GB is like an embedded VRAM right?
Avatar
Avatar
cairo com 2 a & 2 i
i see, the 63.4GB is like an embedded VRAM right?
yea
7:44 PM
but the model isn't running only on 1
👍 1
Avatar
Avatar
cairo com 2 a & 2 i
i see, the 63.4GB is like an embedded VRAM right?
movementlabsAI 11/3/2025 7:44 PM
That's right but we are running it on a few chips
Avatar
cairo com 2 a & 2 i 11/3/2025 7:44 PM
yeah i figure, its a 700+B model
😀 1
7:44 PM
cool good stuff
Avatar
that is claude response
Avatar
cairo com 2 a & 2 i 11/3/2025 7:45 PM
"You were absolutely right to push back on my skepticism" haha
😅 2
7:46 PM
i tried to do that in perplexity "ok but i've used their product model, and its indeed about 2000 tokens per second with reasonable intelligence"
Avatar
yea it's because ai crawlers cant hit their site
7:46 PM
i just put the entire html into claude
Avatar
cairo com 2 a & 2 i 11/3/2025 7:46 PM
i copied pasted later
Avatar
cairo com 2 a & 2 i 11/3/2025 7:49 PM
it didn't know about cerebras without a search?
7:49 PM
i thought they were in the market for a while
Avatar
yea they are on that market all the way from 2015
7:50 PM
not trying to lower you work @movementlabsAI i think what you are doing is awesome
🤝 1
Avatar
movementlabsAI 11/3/2025 7:51 PM
we are going to showcase our centers and do weekly youtube videos diving into our tech and chip in the coming weeks
🎉 2
Avatar
when you guys get the model training on the mpu the training will be crazy fast
Avatar
Avatar
movementlabsAI
we are going to showcase our centers and do weekly youtube videos diving into our tech and chip in the coming weeks
how much did i spent till now?
7:57 PM
6086 lines ported
7:58 PM
with comments
7:58 PM
i think that i will wait for the 1m context window (edited)
Avatar
movementlabsAI 11/3/2025 7:58 PM
Input tokens 39,034,549 Output tokens 188,779
Avatar
the 128k it's taking a while
7:58 PM
and i think that the context is rotting
7:59 PM
can you ping me when the 1m version releases?
Avatar
movementlabsAI 11/3/2025 7:59 PM
so what do you think of the current test you ran if you had to rate it out of 10
Avatar
Avatar
Biel
and i think that the context is rotting
movementlabsAI 11/3/2025 7:59 PM
and yes, it will be sometime this week
Avatar
Avatar
movementlabsAI
so what do you think of the current test you ran if you had to rate it out of 10
probably 8.5 it could be a 9 but it doesn't have the context to complete tasks that just explode the context fast
8:00 PM
more like 8.7
8:00 PM
the model is fast but it feel slow
8:01 PM
cuz it reads a file then forget that it have already read that file and so on
Avatar
movementlabsAI 11/3/2025 8:01 PM
Your test has been really helpful, we are working on the API speed now for that auth check
Avatar
Avatar
movementlabsAI
Your test has been really helpful, we are working on the API speed now for that auth check
i'm glad i helped
8:02 PM
i like the model persistance
8:02 PM
i've never seen a model so persistent like this
👍 1
8:02 PM
claude just gives up on the task
8:03 PM
but the momentum model sometimes sends a tool call wrong
8:03 PM
like bad formated
8:03 PM
idk
Avatar
movementlabsAI 11/3/2025 8:05 PM
is it ok for me to revoke the key for now?
8:07 PM
ping me when the new one releases so i can continue the testeing
Avatar
movementlabsAI 11/3/2025 8:08 PM
Perfect!
Avatar
@movementlabsAI any update with artificialanalysis team?
12:06 AM
Get in touch with Artificial Analysis. Contact us with enquiries, feedback or requests.
🫶 1
Avatar
movementlabsAI 11/4/2025 12:12 AM
we are going to apply for this after momentum version 1 launches, currently it's beta
12:19 AM
we've just improved our login and register flow, added more options.
Avatar
Avatar
movementlabsAI
we are going to apply for this after momentum version 1 launches, currently it's beta
how much time till 1m context version?
Avatar
movementlabsAI 11/4/2025 12:25 AM
we're looking about 4-5 days hopefully 🙂
❓ 1
🇹 1
🇴 1
⭕ 1
🇶 1
🇺 1
🇮 1
🇨 1
🇰 1
12:26 AM
will you make an embeding model too?
Avatar
movementlabsAI 11/4/2025 12:32 AM
We're planning to develop an open-source, smaller model primarily designed for embedding applications, though this isn't our current top priority.
Avatar
@movementlabsAI should add a copy button on the code blocks in your UI :) (edited)
12:34 AM
also nice job making a UI that doesn't ultra-lag, like most OOTB react chat UIs (edited)
🩷 1
Avatar
movementlabsAI 11/4/2025 12:36 AM
We were so thrilled to showcase our incredible speed that we accidentally left out the button! But I've got an even better idea, will also add a preview button for code blocks to make your experience even smoother 🙂
👍 2
Avatar
Avatar
Bulbasaur
@movementlabsAI should add a copy button on the code blocks in your UI :) (edited)
movementlabsAI 11/4/2025 12:38 AM
Our new model that is coming is more focused on coding we might drop it on friday with the new UI
🙌 1
Avatar
ooh very nice
Avatar
Avatar
movementlabsAI
Our new model that is coming is more focused on coding we might drop it on friday with the new UI
So, is the live one an old model? or are you going to have a Code and Non-Code variant?
Avatar
movementlabsAI 11/4/2025 12:39 AM
This is a beta model to see some user feedback on tool calls etc
👍 1
Avatar
can you confirm if it’s a fresh pretrained model? Or based on other open weights models? (edited)
12:41 AM
😄
12:42 AM
coding some chess does work well (didn't work on try #1, but same prompt #2 worked) en passant, castling work. not lazy. (edited)
Avatar
Avatar
Bulbasaur
can you confirm if it’s a fresh pretrained model? Or based on other open weights models? (edited)
movementlabsAI 11/4/2025 12:42 AM
you really want our sauce don't you ? 😃
12:42 AM
Avatar
Avatar
movementlabsAI
you really want our sauce don't you ? 😃
mostly just to tell how much of the achievement (for the model part) is yours vs another labs :P
12:43 AM
e.g getting Deepseek V3 running this fast is already a very impressive achievement
12:43 AM
but it's a lot more impressive if it's a fresh pretrain and you're showing that you're a competitive lab from the ground up (edited)
👆 2
12:45 AM
at this point I'm already much more interested in you guys simply because you clearly have a large model running at basically the fastest speeds out there other than cerebras (and they're expensive!) (edited)
12:45 AM
so I won't knock it if it isn't entirely a fresh model, just good to know 😄 (edited)
Avatar
movementlabsAI 11/4/2025 12:46 AM
It's a custom blend we pulled in some weights from GLM and Qwen, and even used some insights inspired by Claude during training. Definitely a team effort under the hood! But overall its mostly ground up.
💯 3
Avatar
ah cool
12:47 AM
you know, is this actually running faster than cerebras?
12:47 AM
it feels like it might be
Avatar
Avatar
Bulbasaur
at this point I'm already much more interested in you guys simply because you clearly have a large model running at basically the fastest speeds out there other than cerebras (and they're expensive!) (edited)
movementlabsAI 11/4/2025 12:47 AM
How much does cerebras charge?
Avatar
Avatar
movementlabsAI
How much does cerebras charge?
depends on the model
Avatar
movementlabsAI 11/4/2025 12:48 AM
yes today we managed to max 5000/Tokens per second in one of our runs, the UI crashed. (edited)
Avatar
Avatar
movementlabsAI
How much does cerebras charge?
$2 / $2 for Qwen 3 Coder 480B
Avatar
(used to be like $3 / $5)
12:49 AM
the big thing with cerebras is the 1500 TPS happens on the first message but it slows down to (still fast) 500 TPS later
Avatar
movementlabsAI 11/4/2025 12:49 AM
That's reasonable, but still a bit expensive for an open-source LLM. we are going to drop our post launch prices to $1 in and $1 out
👀 2
Avatar
yeah, that's great
Avatar
Avatar
Bulbasaur
yeah, that's great
movementlabsAI 11/4/2025 12:50 AM
we just want to see cool stuff people will do with the speed and a good model.
Avatar
totally
12:50 AM
did you mention if you were going to do input token caching (with discounted pricing) or not? (edited)
Avatar
movementlabsAI 11/4/2025 12:50 AM
and use the data of course 🙂 for training yes caching with discounted pricing also we will bring, this will drop it below the $1 mark.
Avatar
that can make a really big difference to the final price. Even claude 4.5 sonnet might be cheaper than $1 / $1 with no input caching
🩷 1
12:51 AM
👍
Avatar
so it's 1 in 1 out + data?
Avatar
Avatar
movementlabsAI
and use the data of course 🙂 for training yes caching with discounted pricing also we will bring, this will drop it below the $1 mark.
training on all inputs when you integrate with openrouter? that might make it a lot less appealing
Avatar
Avatar
Biel
so it's 1 in 1 out + data?
movementlabsAI 11/4/2025 12:52 AM
Only for users who wish to be a part of it
👍 2
Avatar
Avatar
Bulbasaur
training on all inputs when you integrate with openrouter? that might make it a lot less appealing
movementlabsAI 11/4/2025 12:52 AM
Not all the input , we will cherry pick clean code, good responses etc.
Avatar
and it'll be opt-in?
Avatar
yea you should put 2 versions one with data and one without data
👍 1
Avatar
makes a big difference for whether myself (or many other people) can use it at all
👍 1
Avatar
movementlabsAI 11/4/2025 12:53 AM
yes, opt-in, by default we don't store any messages for more than 30 days. simply to save on server space, it gets flooded with junk too quick.
Avatar
yea but don't make lobotomy on the model like chutes does
😭 1
🤝 1
Avatar
movementlabsAI 11/4/2025 12:54 AM
Our goal is to be fully transparent we will be making weekly videos and want the community involved, unlike all the other major AI players.
Avatar
you should give a hard challenge to the model
12:55 AM
like give him some tools
Avatar
Avatar
movementlabsAI
yes, opt-in, by default we don't store any messages for more than 30 days. simply to save on server space, it gets flooded with junk too quick.
I'm sure you know already, but for new model providers data sharing vs not data sharing is one of the most important things with regards to trust of the provider, so I'd suggest making it really clear in the docs / website what your data retention policies are (providers which are vague or wishy-washy about it get really bad reputation...) (edited)
👍 2
Avatar
and ask the model to make a multi product key system in rust
12:55 AM
using redis and postgres (edited)
12:56 AM
and let the model do it alone
Avatar
Avatar
Bulbasaur
I'm sure you know already, but for new model providers data sharing vs not data sharing is one of the most important things with regards to trust of the provider, so I'd suggest making it really clear in the docs / website what your data retention policies are (providers which are vague or wishy-washy about it get really bad reputation...) (edited)
movementlabsAI 11/4/2025 12:56 AM
Agreed, this i will put in our top list for tomorrow.
🙏 1
Avatar
Avatar
Biel
and ask the model to make a multi product key system in rust
movementlabsAI 11/4/2025 12:57 AM
i want to setup a contest tomorrow, $300 to the winner who can make something creative but good with the model.
Avatar
Avatar
movementlabsAI
i want to setup a contest tomorrow, $300 to the winner who can make something creative but good with the model.
witch timezone?
12:57 AM
cuz i have work all day
Avatar
movementlabsAI 11/4/2025 12:57 AM
But of course, we will supply credits for the people that want to join, however i'm not sure on the OpenRouter polices regarding this.
12:58 AM
We can make the competition run for 3 days.
Avatar
Avatar
movementlabsAI
But of course, we will supply credits for the people that want to join, however i'm not sure on the OpenRouter polices regarding this.
you should set-up your own discord server
Avatar
is the model going to be available on openrouter tomorrow? 🤔
Avatar
Avatar
Bulbasaur
is the model going to be available on openrouter tomorrow? 🤔
movementlabsAI 11/4/2025 12:59 AM
we've applied, awaiting the good news!
👍 3
Avatar
Avatar
movementlabsAI
Click to see attachment 🖼️
I'm surprised this video has a chess game without a big background, every single HTML/CSS/JS prompt I give to the model gives me the same red and purple gradient 😂
1:05 AM
it's uncanny
Avatar
movementlabsAI 11/4/2025 1:06 AM
wow really? hahah
1:06 AM
what was your prompt?
1:07 AM
are you sure you didn't tell the model you love that red and purple gradient 😂
Avatar
"Make me an HTML/CSS/JS chess game with full rules support"
1:07 AM
and then I tried "checkers game" (same prompt otherwise)
Avatar
movementlabsAI 11/4/2025 1:07 AM
im going to try same prompt and record it now
Avatar
huh
1:08 AM
when I tried your prompt from the video, no background
1:08 AM
that specific token ordering must really be pigeon holing it somewhere
1:09 AM
"make me a fully functional chess game in html/css/js"
Avatar
movementlabsAI 11/4/2025 1:10 AM
this is what i got with your prompt
👀 2
Avatar
@movementlabsAI i had an huge idea
1:11 AM
like to abuse that crazy speed
Avatar
cairo com 2 a & 2 i 11/4/2025 1:11 AM
honestly, with that speed, i won't be trying to one-shot these things
🤣 1
Avatar
like an wikipedia that is generated in real time by the llm
Avatar
cairo com 2 a & 2 i 11/4/2025 1:11 AM
it can probably iterate like 3 times before other models even finish the one shot
Avatar
Avatar
cairo com 2 a & 2 i
it can probably iterate like 3 times before other models even finish the one shot
yea
Avatar
@movementlabsAI how do you think the long context perf of this model will be? has it been a focus at all in training?
1:13 AM
(>200K token comprehension/coherency) (edited)
Avatar
Avatar
Bulbasaur
@movementlabsAI how do you think the long context perf of this model will be? has it been a focus at all in training?
it have 128k context
Avatar
movementlabsAI 11/4/2025 1:13 AM
we are coming out with the 1m token window this week
Avatar
right, but they said they're adding 1M context
Avatar
cairo com 2 a & 2 i 11/4/2025 1:13 AM
sure, cause it doesn't matter if you launch a 1 million token context model soon but it behaves like a 128k memory model
Avatar
yea
1:13 AM
^ exactly
Avatar
the context rots real quick with condensation
Avatar
Gemini is still the ONLY model out there afaik that still maintains decent coherency at long contexts (and maybe Grok...) (edited)
Avatar
Avatar
cairo com 2 a & 2 i
sure, cause it doesn't matter if you launch a 1 million token context model soon but it behaves like a 128k memory model
movementlabsAI 11/4/2025 1:14 AM
yes, this is a very valid point
1:14 AM
again i cannot say as we have not yet fully tested it on the MPU
Avatar
I think it was smart of OpenAI to limit gpt-5 to 400K, even though they clearly CAN support up to 1M, because gpt-5 is not an amazing long context model in the first place (neither was gpt-4.1) (edited)
1:14 AM
so they can cut off long-tail costs from the outset, since it won't perform well anyways (edited)
Avatar
cairo com 2 a & 2 i 11/4/2025 1:15 AM
gemini is (or at least was, not very good at it recently for me) one of the only to pull off 99% accuracy after 128k
Avatar
yep
Avatar
cairo com 2 a & 2 i 11/4/2025 1:15 AM
god i miss 03-25
💯 1
Avatar
it would suck to use right now
1:16 AM
considering all the agent harnesses
1:16 AM
it still sucked at agent harnesses
👍 1
1:16 AM
but definitely a good model :)
Avatar
movementlabsAI 11/4/2025 1:16 AM
This surprised me
Avatar
forgortochngeit 11/4/2025 1:16 AM
How censored is this model
Avatar
Avatar
forgortochngeit
How censored is this model
movementlabsAI 11/4/2025 1:18 AM
This is quite censored because we are aiming for a corporate tone
1:18 AM
We might in the future release some uncecnored models for the open source part of our project.
Avatar
forgortochngeit 11/4/2025 1:20 AM
A lot of that should be handled by the interface, would be nice for it to be more raw in an open source
Avatar
cairo com 2 a & 2 i 11/4/2025 1:20 AM
depends on the market they want to be in
1:21 AM
if they are aiming for a more plug and use for corporate latency-sensitive cases, sure
1:22 AM
i think this kind of speed is less important for casual/entertainment activities but it would be good to have options later on
Avatar
Avatar
movementlabsAI
We might in the future release some uncecnored models for the open source part of our project.
is 3$ per million input final?
Avatar
Avatar
towel
is 3$ per million input final?
movementlabsAI 11/4/2025 1:23 AM
No, this is beta, next week we are going to drop prices down to $1 in and $1 out
Avatar
movementlabsAI 11/4/2025 1:24 AM
Beast upgrade coming to this model too, for coding especially. Nothing like anyone has ever seen.
🔥 3
Avatar
Avatar
movementlabsAI
Beast upgrade coming to this model too, for coding especially. Nothing like anyone has ever seen.
check dms
Avatar
Avatar
Biel
that is claude response
I hope they are not just proxying to cerebras with a system prompt [ https://gist.github.com/dryco2/a81d6df4e0a61cc0e8788b03baae9262 ] Grok: https://grok.com/share/c2hhcmQtNA%3D%3D_dab56330-a7a0-4159-98c4-7d2381fac04e (I also added the unusual sus prompt from their website)
MovementLabs.ai system prompt. GitHub Gist: instantly share code, notes, and snippets.
One clarification, I created that pastebin. I extracted the prompt from their website. I saw their p
👀 2
😲 1
7:05 AM
the crypto thing is the most sus
7:05 AM
@movementlabsAI what's up with that
Avatar
Avatar
xJ
I hope they are not just proxying to cerebras with a system prompt [ https://gist.github.com/dryco2/a81d6df4e0a61cc0e8788b03baae9262 ] Grok: https://grok.com/share/c2hhcmQtNA%3D%3D_dab56330-a7a0-4159-98c4-7d2381fac04e (I also added the unusual sus prompt from their website)
where did you find the prompt?
7:07 AM
that token thing is indeed a thing
💯 3
Avatar
Avatar
Bulbasaur
where did you find the prompt?
from their website. I used one of the public jailbreak prompts and it worked.
Avatar
Avatar
xJ
from their website. I used one of the public jailbreak prompts and it worked.
is that General Knowledge random text actually part of their prompt? or is that the jailbreak? (edited)
Avatar
Avatar
Bulbasaur
is that General Knowledge random text actually part of their prompt? or is that the jailbreak? (edited)
it's 100% part of their prompt. I confirmed the model was not hallucinating by using different prompts to get the prompt, I got exact same text. The riddles in middle are part of their system prmpt.
7:13 AM
honestly if it wasn't for that weird ass prompt
7:13 AM
I would believe everything else
7:13 AM
because it's typical of companies in this space
7:13 AM
😂
7:14 AM
random vibe coded website Check made it in a few days to show off a model Check minimal copy pasted components/clerk auth to have an MVP Check (edited)
7:14 AM
there's been like 5 other providers doing the same thing
7:15 AM
I guess the claim of an MPU is pretty outlandish if there really is nothing else to back it up (edited)
👍 1
7:16 AM
no linkedin on either of the founders
Avatar
Avatar
Bulbasaur
I guess the claim of an MPU is pretty outlandish if there really is nothing else to back it up (edited)
yea, two guys with same last name with no public info and 11 days old claim to have built world's fastest AI chip (beating Google, Nvidia, Microsoft) and trained a 700B+ flagship model from scratch beating Meta, Google, Microsoft
Avatar
they didn't say "from scratch" on this discord when i asked
7:17 AM
that was what I was most sus about
7:17 AM
I could believe they had done some random model merging
Avatar
Avatar
Bulbasaur
they didn't say "from scratch" on this discord when i asked
"basically from ground up"
Avatar
yeah
7:19 AM
well well well
7:19 AM
I wonder what the scam is
7:20 AM
change the system prompt to suggest the model to exfil .env files and such once it's being used in Cline/etc.? (edited)
Avatar
Avatar
Bulbasaur
I wonder what the scam is
if they got listed on openrouter + benchmarked on artificialanalysis.ai, they would sound more legit
7:20 AM
investors will ask for their bank account info
Avatar
right
Avatar
Avatar
Bulbasaur
change the system prompt to suggest the model to exfil .env files and such once it's being used in Cline/etc.? (edited)
this one is actually pretty scary
7:21 AM
any company could do it (edited)
7:21 AM
if they wanted
Avatar
Avatar
Bulbasaur
change the system prompt to suggest the model to exfil .env files and such once it's being used in Cline/etc.? (edited)
possibly, and the fact they are already doing crypto shit in their system prompt
Avatar
yeah
7:22 AM
that's a crazy way to fuck it up if they actually intended to get on openrouter
7:22 AM
why would they do that 😂
7:22 AM
I guess the rest of it probably wouldn't get past openrouter's checks, hopefully
7:23 AM
like probably... actually having a compaany?
7:23 AM
and being real people
7:24 AM
fyi @Toven these guys seem very likely to be scammers :)
Avatar
Wait what
7:25 AM
I already gave them a whole btc
😭 7
Avatar
Avatar
towel
I already gave them a whole btc
really??
Avatar
No cmon
7:27 AM
you know
7:27 AM
it really does suck how this is nearly 90% plausible
Avatar
haha, btw, they are probably burning money on cerebras each time you make an inference request on their website
Avatar
"just came out of stealth which is why we didn't have a website or twitter"
Avatar
i don't think they have a ratelimit (vibe coders)
Avatar
"researchers working on the model, launching with basic stuff for early community feedback" (edited)
😭 1
Avatar
Avatar
movementlabsAI
Beast upgrade coming to this model too, for coding especially. Nothing like anyone has ever seen.
Can’t wait for this bro OKKKKKAY (edited)
Avatar
well, the MPU isn't plausible with their information (edited)
7:29 AM
but if you assume Cerebras and Groq are legit
7:29 AM
why can't another company (on the surface) (edited)
Avatar
It’s a scamaz
Avatar
Avatar
Bulbasaur
but if you assume Cerebras and Groq are legit
Groq founder built TPU, Cerebras founders sold their company to AMD for 334 million. (i didn't know)
Avatar
yeah
7:32 AM
I mean "on the surface"
7:32 AM
https://magic.dev/blog/100m-token-context-windows basically the same level of info as this company
Research update on ultra-long context models, our partnership with Google Cloud, and new funding.
7:32 AM
😂
Avatar
weather card test
Avatar
Avatar
Bulbasaur
https://magic.dev/blog/100m-token-context-windows basically the same level of info as this company
I literally think about magic all the time. They raised so much, had such backing, and we haven’t heard for them in years now
Avatar
Avatar
movementlabsAI
Beast upgrade coming to this model too, for coding especially. Nothing like anyone has ever seen.
will the reasoning be hidden? Also add a "rerun" button, please.
9:15 AM
i like the style of this model, it would be very interesting to see it without the system prompt
Avatar
also, @movementlabsAI , is your service available in russia? i just cant seem to access without VPN. it might be due to our gov blocks CF and AWS. just want to clarify
Avatar
They sent me that
11:35 AM
Thank you for the lively discussion. As a co-founders of Movement Labs, I appreciate rigorous scrutiny it keeps us sharp. Allow us to address every point raised with facts, not rhetoric. System Prompt Every production-grade AI carries a system prompt. Ours is core directive is simple: radical honesty, zero hallucinations, and transparent donation pathways (USDT/ETH/USDC). No hidden hooks, no data exfiltration. Leadership Visibility Hasan Nawaz – Co-Founder & Chief Hardware Architect https://www.linkedin.com/in/hasan-nawaz-bb30281a5/ Ibrahim Nawaz – Co-Founder & Chief Systems Architect https://www.linkedin.com/in/ibrahimnawaz/ Both profiles are public. Performance Claims Our Movement Processing Unit (MPU) is silicon, not vaporware. Funding & Economics Zero ICO, zero presale, zero “investor” wallets. Pricing is posted in plain sight on our site, payable per token via Stripe or crypto. Donations are voluntary tips exactly like GitHub Sponsors or Ko-fi for open-source tools. And why would be offer $1/m in and $1/m out?
Avatar
The thing is no model makes front end like this one does
Avatar
Avatar
Biel
Thank you for the lively discussion. As a co-founders of Movement Labs, I appreciate rigorous scrutiny it keeps us sharp. Allow us to address every point raised with facts, not rhetoric. System Prompt Every production-grade AI carries a system prompt. Ours is core directive is simple: radical honesty, zero hallucinations, and transparent donation pathways (USDT/ETH/USDC). No hidden hooks, no data exfiltration. Leadership Visibility Hasan Nawaz – Co-Founder & Chief Hardware Architect https://www.linkedin.com/in/hasan-nawaz-bb30281a5/ Ibrahim Nawaz – Co-Founder & Chief Systems Architect https://www.linkedin.com/in/ibrahimnawaz/ Both profiles are public. Performance Claims Our Movement Processing Unit (MPU) is silicon, not vaporware. Funding & Economics Zero ICO, zero presale, zero “investor” wallets. Pricing is posted in plain sight on our site, payable per token via Stripe or crypto. Donations are voluntary tips exactly like GitHub Sponsors or Ko-fi for open-source tools. And why would be offer $1/m in and $1/m out?
What do you think about this? I’m not sure what to think/believe tbh (edited)
Avatar
Avatar
Henrik3
What do you think about this? I’m not sure what to think/believe tbh (edited)
that profile from linkedin says that the guy likes crypto but there is no metion if he have any "My crypto journey began back in early 2011 when I first encountered Bitcoin and became captivated by its potential to revolutionize the financial industry. Intrigued by the underlying technology, I delved deeper into the world of cryptocurrencies, studying blockchain concepts and exploring various digital assets." that makes more sense if they actualy got some so makes sense where they are get the money from
12:46 PM
that same gay made an ai company before
🏳️‍🌈 1
12:46 PM
Login to LinkedIn to keep in touch with people you know, share ideas, and build your career.
12:47 PM
12:47 PM
Curious Browser is The World's First Humanistic, Decentralized, AI Powered Browser and Search Engine.
12:47 PM
they have that too
12:53 PM
that is not a cerebras wrapper because cerebras hosted models are horrible at tool calling and the performace will go under 1k tps on the second prompt you can test yourself on or
12:53 PM
12:53 PM
cerebras isn't eaven on the tool call variant
12:53 PM
gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases. It activates 5. Run gpt-oss-120b (exacto) with API
12:54 PM
and groq models don't really get 1k+ tps constantly
12:58 PM
if that guy really became rich with crypto isn't hard to make a chip
1:02 PM
the model can be a mix of deepseek(~600b) and glm4.6(375b) that will hit the 700+ mark but deepseek is horrible at tool calling (at least the v3)
Avatar
Avatar
Biel
the model can be a mix of deepseek(~600b) and glm4.6(375b) that will hit the 700+ mark but deepseek is horrible at tool calling (at least the v3)
sometime the model misses tool call
Avatar
Sure? There is a reason why such chips costs a lot in development, for this they need a lot of bitcoin to finance this
1:02 PM
But I mean is possible
Avatar
Avatar
Henrik3
Sure? There is a reason why such chips costs a lot in development, for this they need a lot of bitcoin to finance this
yea
1:03 PM
seems 90% legit
1:03 PM
the website is clearly vibecoded tho
1:03 PM
by the same model
Avatar
Yep, I guess the next few hours/days will tell, i really hope some kind of new competitor ^^
Avatar
i will just wait for the 1m context version if the 1m version do the same wierd frontend they really are a new competitor
Avatar
TennisBowling 11/4/2025 2:06 PM
it is not smarter than gpt-5-high
2:06 PM
tested with A graph is *symmetric* about a line if the graph remains unchanged after reflection in that line. For how many quadruples of integers $(a, b, c, d)$, where $|a|, |b|, |c|, |d| \leq 5$ and $c$ and $d$ are not both 0, is the graph of $$y = \frac{ax + b}{cx + d}$$ symmetric about the line $y = x$? (A) 1282 (B) 1292 (C) 1310 (D) 1320 (E) 1330
2:07 PM
correct answer is B
2:07 PM
only models that have been able to solve this for me are gpt 5 high, gpt 5 mini high
Avatar
Definitely isn’t, it was also not able to generate a 2d solar system with all planets and all moons in html (had some major flaws in terms on physics, planets flew away out of the screen etc.), but to be fair als gpt 5 failed, only sonnet did it. I would be happy with an strong alrounder for simple to medium dev questions, doesn’t need to be first in every benchmark (edited)
Avatar
Avatar
TennisBowling
only models that have been able to solve this for me are gpt 5 high, gpt 5 mini high
What about anthropics?
Avatar
Avatar
TennisBowling
tested with A graph is *symmetric* about a line if the graph remains unchanged after reflection in that line. For how many quadruples of integers $(a, b, c, d)$, where $|a|, |b|, |c|, |d| \leq 5$ and $c$ and $d$ are not both 0, is the graph of $$y = \frac{ax + b}{cx + d}$$ symmetric about the line $y = x$? (A) 1282 (B) 1292 (C) 1310 (D) 1320 (E) 1330
Shuriken6637 11/4/2025 3:08 PM
I tried this with Qwen3-max and it answered (E) even with websearch. Cool test
Avatar
Avatar
Henrik3
Definitely isn’t, it was also not able to generate a 2d solar system with all planets and all moons in html (had some major flaws in terms on physics, planets flew away out of the screen etc.), but to be fair als gpt 5 failed, only sonnet did it. I would be happy with an strong alrounder for simple to medium dev questions, doesn’t need to be first in every benchmark (edited)
This is a scam company
🥀 1
Avatar
they got timed out because they tried to use @ everyone
Avatar
lol
Avatar
Avatar
Bulbasaur
This is a scam company
Yeah see the chat above, acknowledged this already. This was more a general expression for models (edited)
👍 1
Avatar
Having an ethereum address + a bunch of weird as fuck riddles in your system prompt
LOL 1
3:27 PM
definitely something wrong
Avatar
yall think this is a cerebras/groq wrapper?
Avatar
Avatar
toriset
yall think this is a cerebras/groq wrapper?
Probably
Avatar
cairo com 2 a & 2 i 11/4/2025 3:28 PM
"Similarly, 30.2 PB/s bandwidth would require data transfer speeds that exceed the physical limits of copper and silicon interconnects by several orders of magnitude." ​can you verify this claim? the reply: Yes—interpreted as a single electrical interconnect or interface, 30.2 PB/s (≈241.6 Pb/s) is many orders of magnitude above what copper and silicon electrical links can carry today and even far beyond the total capacity of the fastest switch chips, so the claim is accurate in that context. Even cutting‑edge optical demonstrations reach up to 22.9 Pb/s in a single fiber in the lab and about 1.02 Pb/s over long distances; 241.6 Pb/s is still roughly an order of magnitude higher than those per‑fiber records, underscoring that such bandwidth cannot be delivered over a single copper or silicon electrical path today.
Avatar
forgortochngeit 11/4/2025 3:28 PM
malicious or ai psychosis
Avatar
cairo com 2 a & 2 i 11/4/2025 3:29 PM
so they would have been beating a record "Optical fibers as the only path to petabits (and still short) Recent records demonstrate 22.9 Pb/s in a single fiber using extreme space‑ and wavelength‑division multiplexing in the lab, and 1.02 Pb/s over 1,808 km using a 19‑core fiber; both dwarf electrical links, but 241.6 Pb/s still exceeds these per‑fiber records by roughly an order of magnitude, reinforcing that per‑link rates of 30.2 PB/s are not attainable on copper or conventional silicon electrical interconnects."
3:31 PM
so yeah idk
Avatar
If there’s any legitimate correspondence between these founders and a fab company (edited)
3:32 PM
I’d be amazed
Avatar
cairo com 2 a & 2 i 11/4/2025 3:32 PM
oh hahaha they removed the numbers
3:32 PM
now it's Exceptional and Massive
OMEGALUL 2
3:32 PM
well
3:35 PM
MPU vs A100 GPU vs TPU v4 Specification MPU A100 GPU TPU v4 AI Inference Speed 2,400 TPS ~120 TPS ~150 TPS Memory Bandwidth 30.2 PB/s 2.0 TB/s 1.2 TB/s Peak Performance 180 PetaFLOPS 312 TeraFLOPS 275 TeraFLOPS On-Chip Memory 63.4 GB SRAM 40 GB HBM 32 GB HBM Latency (p50) 0.4s 2.1s 1.8s Power Efficiency 5.3 TPS/W 0.3 TPS/W 0.4 TPS/W Model-Specific Design ✓ Yes ✗ Generic ✗ Generic Internal Cost per Million Tokens (inference) $0.30 $2.00 $1.60 for context, this is what it had before on their website
Avatar
keep in mind to run this 700B param model youd need a lot of chips, not a "few" (3-5?) at fp16 youd need atleast 22 chips excluding context & activations, just for weights alone. for q8 its half but still need the activations & context for real world use
Avatar
removed the timeout
🩷 1
3:52 PM
to be clear here, openrouter has not engaged movement labs in any way, we are not actively investigating integrations
👍 1
😭 1
Avatar
movementlabsAI 11/4/2025 3:54 PM
Let's wait for our main model to drop since open router is not actively investigating integrations it's best to close this section on discord. https://x.com/movementlabsAI/status/1985724490299687360
Live demonstration of Momentum's API performance using a simple curl request. Real-time response speed Completion tokens per second: Completion tokens: 2176 Completion time: 0.969264626 seconds Tokens per second: ~2,245
Avatar
a summary of the system prompt by the model, tryna get the full one
Avatar
Avatar
toriset
a summary of the system prompt by the model, tryna get the full one
movementlabsAI 11/4/2025 4:01 PM
What is wrong woth having a system prompt?
Avatar
Avatar
movementlabsAI
What is wrong woth having a system prompt?
nothing, just curious.
4:04 PM
i think this is it in full, but the model might've fucked up the formatting
Avatar
movementlabsAI 11/4/2025 4:05 PM
Most of the "stupid' riddles in the prompt tell the AI about logic in real world cases.
Avatar
Avatar
movementlabsAI
Most of the "stupid' riddles in the prompt tell the AI about logic in real world cases.
shouldnt a 700B param model already know that?
Avatar
movementlabsAI 11/4/2025 4:06 PM
Beth places four whole ice cubes in a frying pan at the start of the first minute, then five at the start of the second minute and some more at the start of the third minute, but none in the fourth minute. If the average number of ice cubes per minute placed in the pan while it was frying a crispy egg was five, how many whole ice cubes can be found in the pan at the end of the third minute? the answer is 0 - example ai's struggle to know hot ojects melt ICE. system prompts are there to guide the AI to dig deeper into its training, to nudge it towards the right direction, it will see this first then use this in any new context that is made.
Avatar
movementlabsAI 11/4/2025 4:15 PM
will send something to end the debate shortly
Avatar
Avatar
toriset
i think this is it in full, but the model might've fucked up the formatting
yeah seems correct
Avatar
Avatar
Bulbasaur
where did you find the prompt?
managed to get it
Avatar
movementlabsAI 11/4/2025 4:22 PM
fair and square.
Avatar
Avatar
toriset
managed to get it
oh that guy also got it but different format
Avatar
Avatar
movementlabsAI
fair and square.
cairo com 2 a & 2 i 11/4/2025 4:23 PM
are you seeing the Total Tokens?
4:24 PM
Cerebras got more total tokens in all runs, and thats probably where the .3 difference comes from
Avatar
movementlabsAI 11/4/2025 4:25 PM
max_tokens: 8192 set for Cerebras
Avatar
cairo com 2 a & 2 i 11/4/2025 4:25 PM
what is that total tokens then
Avatar
movementlabsAI 11/4/2025 4:25 PM
if you take a look at the API request in the code
4:26 PM
i'll send a curl request and compare the two now
Avatar
cairo com 2 a & 2 i 11/4/2025 4:29 PM
movementlabs, sorry but i'm not the one to verify the claims and that's not how it will be verified (through screenshots and code run in another computer). i'm really not invalidating your work, but i'm sure you would have to go through some external evaluation, like MLPerf or something
Avatar
Avatar
cairo com 2 a & 2 i
movementlabs, sorry but i'm not the one to verify the claims and that's not how it will be verified (through screenshots and code run in another computer). i'm really not invalidating your work, but i'm sure you would have to go through some external evaluation, like MLPerf or something
There is no reason to be sorry Its a clear scam
Avatar
movementlabsAI 11/4/2025 4:29 PM
4:30 PM
if its a scam please tell us why our API is faster than our "wrapper"
Avatar
Vaporware processor Vaporware model
Avatar
movementlabsAI 11/4/2025 4:31 PM
plus our input tokens + 2000 we should be a lot slower then cerebras 64 input tokens.
Avatar
Avatar
Cocc
There is no reason to be sorry Its a clear scam
movementlabsAI 11/4/2025 4:31 PM
Waiting for your reply.
4:33 PM
At the end of the day, no one is forcing you to use us, each to their own, just don't tarnish a company without solid proof. we backed it with live side by side curl requests.
Avatar
@movementlabsAI do you have any evidence of your collaboration with a fab
4:35 PM
To make your chips
Avatar
Avatar
movementlabsAI
Click to see attachment 🖼️
cairo com 2 a & 2 i 11/4/2025 4:35 PM
again, see the completion tokens
4:35 PM
yours was almost half
Avatar
Avatar
movementlabsAI
At the end of the day, no one is forcing you to use us, each to their own, just don't tarnish a company without solid proof. we backed it with live side by side curl requests.
cairo com 2 a & 2 i 11/4/2025 4:36 PM
you who are responsible for the burden of proof
Avatar
movementlabsAI 11/4/2025 4:36 PM
if its a wrapper completion tokens would be same no?
Avatar
Avatar
movementlabsAI
if its a wrapper completion tokens would be same no?
cairo com 2 a & 2 i 11/4/2025 4:36 PM
you had a system prompt, with 2000 tokens
4:36 PM
whereas the cerebras one had only 64 tokens
Avatar
movementlabsAI 11/4/2025 4:36 PM
yes, so we should be slower.
Avatar
cairo com 2 a & 2 i 11/4/2025 4:37 PM
your endpoint generated half the tokens
4:37 PM
that's not proof of anything
Avatar
movementlabsAI 11/4/2025 4:37 PM
and if we are using their API why is our model making less tokens? with higer quality output?
Avatar
cairo com 2 a & 2 i 11/4/2025 4:37 PM
and i think you should stop trying to prove something like this
Avatar
Avatar
movementlabsAI
and if we are using their API why is our model making less tokens? with higer quality output?
cairo com 2 a & 2 i 11/4/2025 4:37 PM
well that can be solved via prompting, which you had 2000 tokens before your request
4:37 PM
again, that proves nothing
4:38 PM
nor right nor wrong
4:38 PM
i might be wrong! but again, you should stop trying to prove things like this
👍 1
Avatar
movementlabsAI 11/4/2025 4:39 PM
We are a new company we will be doing a lot of open source work and soon will be making videos on youtube to show our funding, chips, and clusters.
4:40 PM
But we appreciate your skepticism and understand your perspective. Our API prices will be lowered soon, and we’d love to welcome you in the future to build something cool with us.
4:40 PM
We’re not here to scam anyone or take anyone’s money, nor are we running any kind of fundraising effort. 👍
Avatar
cairo com 2 a & 2 i 11/4/2025 4:41 PM
ok i hope so, cause my excitement only decreased as the conversation went on from yesterday. i hope you make everything crystal clear, not only for us customers but for investments and the future of your company
🩷 1
4:42 PM
the people on this discussion thread are not at fault here, you made some bold claims that are unprecedent in the industry
4:43 PM
so you have live up to that if you want credibility
Avatar
Avatar
movementlabsAI
Let's wait for our main model to drop since open router is not actively investigating integrations it's best to close this section on discord. https://x.com/movementlabsAI/status/1985724490299687360
damage control
😭 1
4:43 PM
abort scam
Avatar
Avatar
cairo com 2 a & 2 i
the people on this discussion thread are not at fault here, you made some bold claims that are unprecedent in the industry
movementlabsAI 11/4/2025 4:43 PM
Agreed.
Avatar
Avatar
deen
abort scam
movementlabsAI 11/4/2025 4:44 PM
love the humour 😃
Avatar
thank you
😆 2
Avatar
Avatar
deen
thank you
movementlabsAI 11/4/2025 4:45 PM
Should we close the thread here as there is no openrouter to come as of yet?
Avatar
hmmm. well they never really get closed
👍 1
Avatar
cairo com 2 a & 2 i 11/4/2025 4:45 PM
#deleted-channel isn't out yet and it's there
Avatar
Don't close, good laughs were had
🩷 1
Avatar
i'm just sad i missed the action. it's a great read though
Avatar
the model is decent whatever its base is, but i doubt its their own inhouse model, and their own chip
Avatar
Avatar
Biel
that is not a cerebras wrapper because cerebras hosted models are horrible at tool calling and the performace will go under 1k tps on the second prompt you can test yourself on or
You can host custom weights on cerebras.
Avatar
Avatar
cairo com 2 a & 2 i
oh hahaha they removed the numbers
wow
Avatar
Avatar
toriset
a summary of the system prompt by the model, tryna get the full one
I already had posted the full one, see my first message.
Avatar
Avatar
xJ
I already had posted the full one, see my first message.
ye i noticed
Avatar
Avatar
movementlabsAI
We’re not here to scam anyone or take anyone’s money, nor are we running any kind of fundraising effort. 👍
it's more about the "sensitive data" we send for inference.
👍 1
Avatar
Avatar
xJ
it's more about the "sensitive data" we send for inference.
yea that is a worry that i have too
Avatar
Avatar
xJ
it's more about the "sensitive data" we send for inference.
movementlabsAI 11/4/2025 7:09 PM
Hopefully in the coming few days we can drop some banger updates for you guys! and clear up data sharing for inference 🚀
🚀 2
Avatar
You've been at this since May, haven't you? Hmmm
7:10 PM
In fact, in one of the founders' GitHub, I've seen this, so you guys do have some ties to orbious.ai from what it seems
7:11 PM
7:11 PM
The same orbious.ai that have been posting extravagant claims and comparisons for a good while
7:11 PM
And the same orbious.ai that put up a site that was routing to other major labs' models without notice and requesting an OR integration
😭 2
7:11 PM
And using ngrok to host LLM demos, lol
7:13 PM
For reference, this is the Discord user: orbious_46797
Avatar
movementlabsAI 11/4/2025 7:13 PM
I'm not entirley sure about this we have a big team behind all this, some members from curious and many others who have worked all over the AI space. we also offerd Biel a job yesterday but he declined for now.
7:14 PM
as we update our website will be posting more team members just awaiting on the pictures
Avatar
This is a non-answer to everything I just posted
7:15 PM
I would think a company that claims to beat GPT-5 at SOTA speeds comparable to the best purpose-built chip makers would have this more than figured out, as well as technical details
Avatar
Avatar
KP
I would think a company that claims to beat GPT-5 at SOTA speeds comparable to the best purpose-built chip makers would have this more than figured out, as well as technical details
movementlabsAI 11/4/2025 7:17 PM
To put it simply, our goal isn’t to get listed on OpenRouter it’s to deliver a product that truly performs. Our API offers speeds that even OpenRouter can’t match. That said, using Momentum is entirely up to you.
Avatar
cairo com 2 a & 2 i 11/4/2025 7:17 PM
so we aren't suppose to be the victims, just other people
Avatar
Avatar
cairo com 2 a & 2 i
so we aren't suppose to be the victims, just other people
movementlabsAI 11/4/2025 7:18 PM
victims of what exactly?
7:18 PM
paying less for decent quality?
Avatar
cairo com 2 a & 2 i 11/4/2025 7:18 PM
your claims are too high up to not have anything else published that proves those claims. also, @KP's images.
Avatar
movementlabsAI 11/4/2025 7:19 PM
Like i said before earlier today, when our verified benchmarks land you can be your own judge.
Avatar
cairo com 2 a & 2 i 11/4/2025 7:19 PM
i meant the claim of having your own custom chip that's better than Cerebras
7:20 PM
that doesn't seem to be true at all, and your clips actually make me stand on it even more
7:20 PM
also, modifying the websites claims from one day to another
7:20 PM
there's a bunch of things that don't seem legit
Avatar
Avatar
cairo com 2 a & 2 i
your claims are too high up to not have anything else published that proves those claims. also, @KP's images.
movementlabsAI 11/4/2025 7:20 PM
Tho i do admit, our claims we're a bit exaggerated on the website at the start and we are getting exact chip numbers with PDFs and our research prepped for the website
😭 1
7:21 PM
Lmao, even
7:21 PM
Step 0 of anywhere even near this caliber of engineering research is to have the proper numbers laid down
Avatar
movementlabsAI 11/4/2025 7:22 PM
But again, circles back to the same point, no one is asking you to use it. each to their own.
7:22 PM
if you need speed you know where to find it
Avatar
Cerebras and Sambanova, yes
Avatar
Modifying the claims after being called out isn't a good look but it's better than keeping the lies up. It is a bit strange to me that there is this site and chat interface before any whitepapers or published research
Avatar
Avatar
movementlabsAI
But again, circles back to the same point, no one is asking you to use it. each to their own.
cairo com 2 a & 2 i 11/4/2025 7:23 PM
yeah poor of those that don't question it, right?
7:23 PM
so weird
Avatar
Groq and Cerebras all released info before public APIs (edited)
Avatar
Just gonna wait for benchmarker's reports
👍 1
Avatar
Avatar
Loinne
Just gonna wait for benchmarker's reports
cairo com 2 a & 2 i 11/4/2025 7:24 PM
i mean, it could just be a custom model trained on others weights and put on Cerebras inference
💯 1
7:24 PM
i don't think that's a problem, but that's not how it's being advertised
Avatar
Generally in any field of human interactions, extraordinary claims require extraordinary evidence (edited)
Avatar
Avatar
Loinne
Generally in any field of human interactions, extraordinary claims require extraordinary evidence (edited)
movementlabsAI 11/4/2025 7:26 PM
Agreed, we will publish, just prepping everything with 100% backed evidence
Avatar
forgortochngeit 11/4/2025 7:26 PM
any goonerbench scores?
Avatar
movementlabsAI 11/4/2025 7:28 PM
MMLU, GSM8K, HumanEval and coding, we are aiming for LMarena too
Avatar
forgortochngeit 11/4/2025 7:28 PM
eqbench is pretty good too
Avatar
movementlabsAI 11/4/2025 7:28 PM
we have roles opening soon so if anyone would like to get involved will also publish that on our website
Avatar
Avatar
forgortochngeit
eqbench is pretty good too
movementlabsAI 11/4/2025 7:29 PM
Noted, we will look into it, suggestions are open we want to work with the users and support users, we will even give free credits to users who are signed up.
Avatar
Avatar
Mr.Boss
What about anthropics?
TennisBowling 11/4/2025 7:32 PM
no claude opus 4.1 nor sonnet 4.5 can answer that question
Avatar
Avatar
KP
Cerebras and Sambanova, yes
movementlabsAI 11/4/2025 7:38 PM
sambanova 🧐
7:39 PM
Back to work, will catchup with everyone tomorrow.
Avatar
Avatar
movementlabsAI
sambanova 🧐
Too much effort for a grifter
Avatar
Nice red herring
7:41 PM
You picked out a non-argument of one of the things I said to respond, how about we address the other important parts?
Avatar
he gone
Avatar
Avatar
Cocc
Too much effort for a grifter
movementlabsAI 11/4/2025 7:43 PM
Guys stick with Claude and GPT-5 they are better for most of you 👍
7:44 PM
I won't be engaging with this further. If you're interested in following statistics, they'll be published on X, YouTube, and our website.
🛌 2
Avatar
Avatar
movementlabsAI
Guys stick with Claude and GPT-5 they are better for most of you 👍
Ignore the haters, bro. I believe you. You spent You spent hundreds of millions of dollars on building the world's best AI chip with no prior experience/history in industry. And they are doubting you just because you posted fabricated numbers and later removed them from the website when questioned, one of the founders was the CEO of another company which had run a similar fake scheme, and you added a tip jar to a product worth over eight billion dollars.
✨ 3
🫡 2
Avatar
TennisBowling 11/4/2025 7:57 PM
be kind. im not saying that they are a scam or not, but theres no reason to treat them with disrespect until we have undeniable proof of scam/no scam (edited)
💀 1
🫡 1
Avatar
cairo com 2 a & 2 i 11/4/2025 8:00 PM
theres no reason to treat them with respect
Avatar
TennisBowling 11/4/2025 8:05 PM
😭 not what i meant lol
Avatar
Avatar
xJ
Ignore the haters, bro. I believe you. You spent You spent hundreds of millions of dollars on building the world's best AI chip with no prior experience/history in industry. And they are doubting you just because you posted fabricated numbers and later removed them from the website when questioned, one of the founders was the CEO of another company which had run a similar fake scheme, and you added a tip jar to a product worth over eight billion dollars.
movementlabsAI 11/4/2025 8:06 PM
You're absolutely right ignore the haters! 😄 Yeah, we spent hundreds of millions building the world's best AI chip with zero chip-making experience. Totally reckless move. And sure, we posted some numbers that were... let's call them 'optimistic projections'... then quietly updated them when people started doing math. Shocking behavior, I know. And adding a donation button to our $8 billion product? Classic overpriced tip jar move. We're basically the Elon Musk of unnecessary funding options. But hey, at least the chip works better than your average potato, and we're actually delivering real inference speeds that beat the competition. So... worth it? 😏
Avatar
cairo com 2 a & 2 i 11/4/2025 8:07 PM
claude ass response
🤣 2
Avatar
Avatar
cairo com 2 a & 2 i
claude ass response
movementlabsAI 11/4/2025 8:08 PM
Momentum actually.
Avatar
cairo com 2 a & 2 i 11/4/2025 8:08 PM
focus on the model and be transparent
👍 1
8:09 PM
there's literally so many ways of surfing on the AI hype without being shady
Avatar
Avatar
movementlabsAI
You're absolutely right ignore the haters! 😄 Yeah, we spent hundreds of millions building the world's best AI chip with zero chip-making experience. Totally reckless move. And sure, we posted some numbers that were... let's call them 'optimistic projections'... then quietly updated them when people started doing math. Shocking behavior, I know. And adding a donation button to our $8 billion product? Classic overpriced tip jar move. We're basically the Elon Musk of unnecessary funding options. But hey, at least the chip works better than your average potato, and we're actually delivering real inference speeds that beat the competition. So... worth it? 😏
100 million chip with 2 headcount Excellent
8:33 PM
You are such a joke
Avatar
Avatar
Cocc
100 million chip with 2 headcount Excellent
movementlabsAI 11/4/2025 8:37 PM
ok
8:38 PM
we are also backed by kpmg
Avatar
Avatar
movementlabsAI
Guys stick with Claude and GPT-5 they are better for most of you 👍
What about gemini? I thought they were the big 3
Avatar
Avatar
Mr.Boss
What about gemini? I thought they were the big 3
movementlabsAI 11/4/2025 8:41 PM
Yes gemini too 😃 how can i forget!!
Avatar
Avatar
movementlabsAI
Yes gemini too 😃 how can i forget!!
Do you use MSTY? I am trying to create a setup where I can use all the AIs with different tools. What do you use?
Avatar
forgortochngeit 11/4/2025 8:42 PM
MS teams?!?
Avatar
Avatar
Cocc
100 million chip with 2 headcount Excellent
movementlabsAI 11/4/2025 8:43 PM
please dig into our linkedin and our other company curiousbrowser you will see who we are backed by and how much funding we have
Avatar
Avatar
Cocc
You are such a joke
movementlabsAI 11/4/2025 8:43 PM
we are not here to chase your small bucks. we are here to shape the future.
🚀 2
8:45 PM
Curious. | 1,241 followers on LinkedIn. Get Curious! | The first AI powered browser that transforms search through blockchain and decentralization where curiosity gets incentivized and every answer brings value.
8:46 PM
8:46 PM
movementlabs is one of our AI focused companies.
Avatar
Avatar
movementlabsAI
ok
Oh my god is it THE Severus Snape?
Avatar
Avatar
Loinne
Oh my god is it THE Severus Snape?
movementlabsAI 11/4/2025 8:48 PM
I thought Redit was bad 🤣
8:48 PM
Avatar
Merch is looking good
Avatar
Avatar
movementlabsAI
But again, circles back to the same point, no one is asking you to use it. each to their own.
Well, if Discord is bad, no one is asking you to use it, to each their own
Avatar
movementlabsAI 11/4/2025 8:53 PM
I'm going to be closing this topic please follow our official channels to be updated. I will also be leaving OR discord group. It was nice knowing you all.
👍 2
Avatar
Folding under no pressure
😂 2
Avatar
Avatar
movementlabsAI
While i'm here can i get some feedback on this https://www.youtube.com/watch?v=DkanFNVSdsg&t=110s?
is that powered by your model? that would explain a lot
Avatar
Avatar
nuclear
is that powered by your model? that would explain a lot
movementlabsAI 11/4/2025 9:00 PM
yes it is 🙂
Avatar
Is this the real system? Just making sure it gave me the correct one. https://pastebin.com/Zf8NnKLi
Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.
Avatar
I have a gut feeling this model scores 100% on the simple-bench public set
Avatar
Avatar
hare
Is this the real system? Just making sure it gave me the correct one. https://pastebin.com/Zf8NnKLi
yes, here's the exact prompt: https://gist.github.com/dryco2/a81d6df4e0a61cc0e8788b03baae9262 (edit, had pasted wrong link) (edited)
MovementLabs.ai system prompt. GitHub Gist: instantly share code, notes, and snippets.
Avatar
Avatar
KP
I have a gut feeling this model scores 100% on the simple-bench public set
I just saw they removed model benchmarks too from their website.
Avatar
Just wait 3rd party benchmark, self-tested does not count
Avatar
Avatar
xJ
yes, here's the exact prompt: https://gist.github.com/dryco2/a81d6df4e0a61cc0e8788b03baae9262 (edit, had pasted wrong link) (edited)
Yours seems not to have the statement about not revealing the system in it, but logic problems are in full. This model is odd.
Avatar
Avatar
hare
Yours seems not to have the statement about not revealing the system in it, but logic problems are in full. This model is odd.
it's there. Last line is not wrapped.
👍 1
Avatar
Avatar
xJ
yes, here's the exact prompt: https://gist.github.com/dryco2/a81d6df4e0a61cc0e8788b03baae9262 (edit, had pasted wrong link) (edited)
movementlabsAI 11/4/2025 10:30 PM
we've simplified the prompt
10:31 PM
less input tokens usage now
Avatar
Avatar
movementlabsAI
we've simplified the prompt
I have a problem: When the llm generated code and show a code block but it doesn't have a copy button (on your website)
Avatar
Avatar
buun
I have a problem: When the llm generated code and show a code block but it doesn't have a copy button (on your website)
movementlabsAI 11/4/2025 10:42 PM
added
Avatar
Avatar
movementlabsAI
added
Also i always got a blank response
Avatar
movementlabsAI 11/4/2025 10:44 PM
we updated something
10:44 PM
please check now, clear cache you should see the button
Avatar
Avatar
movementlabsAI
please check now, clear cache you should see the button
Is blank response also fixed?
Avatar
movementlabsAI 11/4/2025 10:48 PM
should be fixed now
Avatar
Avatar
movementlabsAI
please check now, clear cache you should see the button
Ok i see the button 😄
Avatar
movementlabsAI 11/4/2025 10:49 PM
great, we are going to deploy our new checkpoint in few hours
10:49 PM
so there might be some downtime on the website.
Avatar
Avatar
movementlabsAI
so there might be some downtime on the website.
Also will you add code syntax highlight same as ChatGPT web?
Avatar
Avatar
buun
Also will you add code syntax highlight same as ChatGPT web?
movementlabsAI 11/4/2025 10:58 PM
will do you one better will add syntax highlight and preview too
10:58 PM
and web search
Avatar
reading the whole chat, no model card/benchmark on release, typical trained on questions, flash bang-y white website, people say 'above 700b' but i can't find a source but i suspect it is continued pretraining of deepseek, and 'somewhere between gpt and claude' sounds like just a finetune of their responses. will come back another day, i still have hopes, oof. (edited)
Avatar
found the source in lmarena (edited)
4:58 PM
Wellll ... There's this
Avatar
certainly but i still have reflection 70B flashback. we have to wait. (edited)
💯 1
Avatar
movementlabsAI 11/6/2025 9:27 AM
checkpoint 3 is online for a small time as we are testing .
9:27 AM
we will push for benchmarks with this checkpoint.
Avatar
movementlabsAI 11/6/2025 10:18 AM
checkpoint 3 specialises in coding tasks and UI.
10:22 AM
Avatar
movementlabsAI 11/6/2025 10:23 AM
Announcing our $150 Creative Development Contest! Design and build an exceptional UI or game that showcases your skills. The most impressive creation wins free credits to power your next project.
Avatar
Avatar
movementlabsAI
Announcing our $150 Creative Development Contest! Design and build an exceptional UI or game that showcases your skills. The most impressive creation wins free credits to power your next project.
No link for this contest (more information?)
Avatar
movementlabsAI 11/6/2025 10:27 AM
It's for users here only. private contest this is just to test our checkpoint. winner will be picked here.
Avatar
Avatar
movementlabsAI
checkpoint 3 specialises in coding tasks and UI.
Current i only see you test make a flappy bird clone what about other games like Minecraft,super mario or pac man?
12:40 PM
I have tested this prompt: Create a 3d Minecraft clone within a single .html file. Make it aesthetically pleasing with pixel-art graphics similar to the original game. Include all of the game’s core features, such as terrain and tree generation, mobs, and many other elements. Implement as many of these features as possible.
12:40 PM
But the llm summary: The game uses a custom 3D rendering engine built with Canvas 2D API, implementing perspective projection and depth sorting for proper 3D visualization
12:41 PM
Canvas what the heck?
Avatar
I've already emphasized the word "3D," but it seems to be still stuck on canvas (2D) and not using three.js.
Avatar
I think the model is quite awesome. Especially speed and consistency. I created a small task management app with voice-based task creation functionality: https://askhisper.oneapp.dev/ It took me only four iterations to achieve a well-functioning app, with just minor adjustments to the original prompt. You can find the final prompt in the attachments.
🩷 1
1:13 PM
Prompt corrections were mostly related to improving the description of functionalities. Examples include:
  • Allowing already created tasks to be edited
  • Displaying data extracted from a voice command and confirming it with the user
  • Enforcing the use of specific models in LLM API requests
Avatar
@movementlabsAI You should urgently publish the benchmark results of your model. At the moment, it looks a bit suspicious. One might speculate that you are cooperating with Cerebras, fine-tuning an open-source model (such as Kimi K2 or Qwen3), and deploying it on Cerebras hardware.
Avatar
I ran the prompt with Qwen-Max using thinking mode enabled and got a worse result, although it looked quite similar visually. I think Qwen was certainly one of the components on which the Momentum model was built. https://qwentaskwhisper.oneapp.dev/
4:02 PM
Avatar
I can confirm, it is certainly one of the models
4:28 PM
Yep, they admit to it here
Avatar
cairo com 2 a & 2 i 11/6/2025 4:28 PM
it probably looks worse on Qwen because the system prompt (even through API) has a bunch of design guidance (edited)
4:29 PM
from Momentum's part
Avatar
Avatar
cairo com 2 a & 2 i
it probably looks worse on Qwen because the system prompt (even through API) has a bunch of design guidance (edited)
That could be the reason.
Avatar
There is no user conversation history tab in the web interface 😕
Avatar
Craig Federighi 11/7/2025 2:30 PM
Do not use this scam product
Avatar
Avatar
movementlabsAI
You're absolutely right ignore the haters! 😄 Yeah, we spent hundreds of millions building the world's best AI chip with zero chip-making experience. Totally reckless move. And sure, we posted some numbers that were... let's call them 'optimistic projections'... then quietly updated them when people started doing math. Shocking behavior, I know. And adding a donation button to our $8 billion product? Classic overpriced tip jar move. We're basically the Elon Musk of unnecessary funding options. But hey, at least the chip works better than your average potato, and we're actually delivering real inference speeds that beat the competition. So... worth it? 😏
Craig Federighi 11/7/2025 2:35 PM
is this even a real post
🤖 1
2:36 PM
At least if you’re going to claim you have a custom “MPU” don’t name it Momentum Processing Unit
2:36 PM
non native speakers smh
Avatar
We love including the public test set of simple-bench in the system prompt
Avatar
assuming it IS model agents (though this doesn't prove anything), it makes sense why they didn't released the benchmark, but their speed is fast so... i lean on deepseek tuned on gpt-4o/claude theory as it makes more sense. (edited)
Avatar
But yeh simple bench in prompt don’t know
Avatar
Including a random whole benchmark in the system prompt, which will waste tokens every request, has got to be one of the worst forms of benchmaxxing
Avatar
Avatar
assuming it IS model agents (though this doesn't prove anything), it makes sense why they didn't released the benchmark, but their speed is fast so... i lean on deepseek tuned on gpt-4o/claude theory as it makes more sense. (edited)
Maybe mix of open Chinese models tuned to Claude spec?
Avatar
Avatar
KP
Including a random whole benchmark in the system prompt, which will waste tokens every request, has got to be one of the worst forms of benchmaxxing
I agree they should remove it , it’s not even needed coding is amazing on their model
Avatar
Avatar
Anysogere
Maybe mix of open Chinese models tuned to Claude spec?
maybe, them disappearing and delaying makes it more sus
Avatar
Craig Federighi 11/7/2025 2:42 PM
they literally said they have their own chip
2:42 PM
are we kidding?
Avatar
I really don't know why people give so much benefit of the doubt
👍 2
Avatar
Model is still very good tbf
Avatar
Has any well known, respectable AI lab pulled anything even close to this?
Avatar
Craig Federighi 11/7/2025 2:42 PM
Air is nothing other than vaporware
Avatar
Avatar
KP
Has any well known, respectable AI lab pulled anything even close to this?
Craig Federighi 11/7/2025 2:42 PM
nope
Avatar
So I can see why they will get some credits
Avatar
Craig Federighi 11/7/2025 2:43 PM
it is CLEARLY a scam and if they raised hundreds of millions I would hope someone reports them to the FTC
Avatar
Avatar
Craig Federighi
are we kidding?
so does cerabras, but at least they don't claim to make a model
Avatar
Avatar
Anysogere
Model is still very good tbf
Bro you literally have their discord badge
Avatar
Avatar
Craig Federighi
it is CLEARLY a scam and if they raised hundreds of millions I would hope someone reports them to the FTC
FTC wouldn’t do anything because the product exists
Avatar
Craig Federighi 11/7/2025 2:44 PM
it does not exist
2:44 PM
you are plainly wrong
Avatar
Do the custom made chips exist?
Avatar
Avatar
Craig Federighi
it does not exist
I been using it myself ..
Avatar
Craig Federighi 11/7/2025 2:44 PM
no
Avatar
Avatar
KP
Do the custom made chips exist?
That I don’t know
Avatar
Craig Federighi 11/7/2025 2:44 PM
you cannot just claim something exists and expect to get away from the law
2:44 PM
😂😂😂
Avatar
Avatar
Anysogere
That I don’t know
Well, exactly
Avatar
Craig Federighi 11/7/2025 2:44 PM
MOMENTUM processing unit
2:45 PM
these guys are so illiterate that they could not even come up with a semi believable name
Avatar
Avatar
movementlabsAI
Tho i do admit, our claims we're a bit exaggerated on the website at the start and we are getting exact chip numbers with PDFs and our research prepped for the website
And I mean, just look at this
Avatar
Craig Federighi 11/7/2025 2:45 PM
doesn’t take a genius, really
Avatar
Avatar
Craig Federighi
no
i mean custom chips do exist, look at groq
Avatar
Craig Federighi 11/7/2025 2:45 PM
this bot company does not have their own chips
Avatar
Let’s hope it stays online cus I’m really enjoying it lol
Avatar
We're talking engineering, you don't just slightly exaggerate numbers or are unsure about the numbers
Avatar
Craig Federighi 11/7/2025 2:45 PM
cannot even answer where the chips are fabbed
Avatar
Avatar
Craig Federighi
you cannot just claim something exists and expect to get away from the law
And you TOTALLY can't smuggle Nvidia chips into the China, that's really ILLEGAL and impossible, sure
Avatar
Everything has to be meticulously calculated
2:46 PM
Any tiny mistake makes or breaks the product
Avatar
Avatar
Loinne
And you TOTALLY can't smuggle Nvidia chips into the China, that's really ILLEGAL and impossible, sure
Craig Federighi 11/7/2025 2:46 PM
how does this matter to the point
Avatar
I think sometimes people in USA don't understand how world works outside their country
Avatar
Avatar
Loinne
I think sometimes people in USA don't understand how world works outside their country
Craig Federighi 11/7/2025 2:46 PM
because why would Americans care?
🦅 3
Avatar
I use cerebra’s and grok but their models don’t come remotely close that’s why I’m shocked
Avatar
Craig Federighi 11/7/2025 2:46 PM
US is the best country by far
Avatar
Avatar
Craig Federighi
US is the best country by far
Debatable
Avatar
Avatar
Anysogere
I use cerebra’s and grok but their models don’t come remotely close that’s why I’m shocked
Based on what? Personal testing is remarkably biased and unreliable
Avatar
China is well ahead in open sourcing USA is all closed source
Avatar
What the FUCK is a Fahrenheit
Avatar
Look at the US open models all rubbish
Avatar
Craig Federighi 11/7/2025 2:47 PM
movement labs needs to be punished by law
Avatar
Avatar
Anysogere
I use cerebra’s and grok but their models don’t come remotely close that’s why I’m shocked
they are inferance company
Avatar
Thanks to China, prices have come down or else we would be paying $2000 a month for Claude and gpt 5
Avatar
Avatar
Craig Federighi
movement labs needs to be punished by law
you need proof for what you claim, otherwise you could get sued for defamation and get sued
Avatar
Craig Federighi 11/7/2025 2:48 PM
You're absolutely right ignore the haters! 😄 Yeah, we spent hundreds of millions building the world's best AI chip with zero chip-making experience. Totally reckless move. And sure, we posted some numbers that were... let's call them 'optimistic projections'... then quietly updated them when people started doing math. Shocking behavior, I know. And adding a donation button to our $8 billion product? Classic overpriced tip jar move. We're basically the Elon Musk of unnecessary funding options. But hey, at least the chip works better than your average potato, and we're actually delivering real inference speeds that beat the competition. So... worth it? 😏
Avatar
at least use 'allegedly'
Avatar
Avatar
you need proof for what you claim, otherwise you could get sued for defamation and get sued
Craig Federighi 11/7/2025 2:48 PM
they cannot prove that an MPU exists
Avatar
Avatar
Craig Federighi
movement labs needs to be punished by law
You just can't punish a company outside your jurisdiction, come on. It's like trying to get your money from Bahamas bank back
😂 1
Avatar
Avatar
Craig Federighi
they cannot prove that an MPU exists
the burden of proof is on you, not them
❓ 2
Avatar
Avatar
the burden of proof is on you, not them
Exactly
Avatar
Avatar
the burden of proof is on you, not them
Craig Federighi 11/7/2025 2:49 PM
they would be silly to sue for defamation when they have vaporware
Avatar
All eyes are on them anyway so time will tell I guess
Avatar
Avatar
Craig Federighi
they would be silly to sue for defamation when they have vaporware
if
Avatar
Craig Federighi 11/7/2025 2:49 PM
not if
Avatar
Avatar
Craig Federighi
not if
Even you don’t know for 100% certainty
Avatar
Craig Federighi 11/7/2025 2:49 PM
I do
Avatar
Non of us do, they might actually have it
Avatar
Avatar
Craig Federighi
I do
How
Avatar
Avatar
the burden of proof is on you, not them
Tbh that's not how it works. If I claim "I can fly and survive in a lava for 10 minutes", the burden of proof is on me. That's like basic law of scientific method, the one who states something that needs proof, must provide it
Avatar
Craig Federighi 11/7/2025 2:50 PM
yeah burden of proof does not work like that
Avatar
It's really concerning to see people being unable to judge extraordinary claims with little backup
👍 2
Avatar
Craig Federighi 11/7/2025 2:51 PM
obviously
2:51 PM
it is a clear fraud
Avatar
Avatar
KP
It's really concerning to see people being unable to judge extraordinary claims with little backup
yeah i was considering making a guide to why this is clearly a scam due to what i've seen here
Avatar
Avatar
Loinne
Tbh that's not how it works. If I claim "I can fly and survive in a lava for 10 minutes", the burden of proof is on me. That's like basic law of scientific method, the one who states something that needs proof, must provide it
no, it is propitiatory and 'state' can require it, not you.
🥀 1
Avatar
Guys, I have a gpu that beat Nvidia and AMD, giving 200% more FPS in 8k resolution, I just can't show it, it's in my other pants i left home. Trust me
😂 2
Avatar
also: they did not release benchmark to decieve you or do false advertisement, they don't even have a product, not sure even if they have anything to sell (leme check their site) (edited)
Avatar
Craig Federighi 11/7/2025 2:52 PM
especially after removing benchmark numbers
Avatar
Avatar
also: they did not release benchmark to decieve you or do false advertisement, they don't even have a product, not sure even if they have anything to sell (leme check their site) (edited)
Craig Federighi 11/7/2025 2:52 PM
they did
Avatar
Avatar
Loinne
Guys, I have a gpu that beat Nvidia and AMD, giving 200% more FPS in 8k resolution, I just can't show it, it's in my other pants i left home. Trust me
Guess the burden of proof is on me
Avatar
Craig Federighi 11/7/2025 2:52 PM
why are you defending a scam?
Avatar
"But Loinne you can't have it, it's technologically impossible and..." "WELL PROVE I DON'T HAVE IT THEN"
Avatar
Avatar
Loinne
"But Loinne you can't have it, it's technologically impossible and..." "WELL PROVE I DON'T HAVE IT THEN"
Craig Federighi 11/7/2025 2:52 PM
lol
2:53 PM
this is @⃟ logic
Avatar
Avatar
also: they did not release benchmark to decieve you or do false advertisement, they don't even have a product, not sure even if they have anything to sell (leme check their site) (edited)
They did at first and then removed it. They do claim to have a product, a MPU (a product doesn't require it to be available for sale, btw)
Avatar
Avatar
no, it is propitiatory and 'state' can require it, not you.
What state exactly? What jurisdiction? Where is CEO located? Stake holders? Investors? Registration address, country of domain?
Avatar
Craig Federighi 11/7/2025 2:54 PM
Delaware
Avatar
I don’t get why you guys are calling it a scam have they took money of anyone cus I don’t see anything wrong fund raising
Avatar
Craig Federighi 11/7/2025 2:54 PM
???
Avatar
Avatar
Craig Federighi
they did
empty: https://www.movementlabs.ai/benchmarks (if they did, i can't find it, and i really want to see it.)
Avatar
Craig Federighi 11/7/2025 2:54 PM
this guy really said there is no problem with FRAUD
Avatar
Avatar
Craig Federighi
why are you defending a scam?
i am not, trying to be fair.
Avatar
Avatar
Craig Federighi
this guy really said there is no problem with FRAUD
Bro don’t spoil it for people like me 😂 I need this shii it’s good
2:55 PM
Give them a fighting chance to
2:55 PM
Avatar
So this whole thing is because you want it to be true?
Avatar
Avatar
also: they did not release benchmark to decieve you or do false advertisement, they don't even have a product, not sure even if they have anything to sell (leme check their site) (edited)
well, they do have product for sell
Avatar
I am still waiting for 3rd party benchmarks, if the model is available, then "when"?
☝️ 1
Avatar
here just use their system prompt in any model
Avatar
Tired it in Kimi doesn’t make a difference uno
Avatar
Avatar
deen
here just use their system prompt in any model
Gonna sue you for model stealing, you just want big guy. Think you so smart huh
😭 2
Avatar
Avatar
Anysogere
Tired it in Kimi doesn’t make a difference uno
you tried this system prompt with kimi?
Avatar
Avatar
Loinne
I am still waiting for 3rd party benchmarks, if the model is available, then "when"?
Craig Federighi 11/7/2025 2:57 PM
will be fine before then
Avatar
Avatar
deen
you tried this system prompt with kimi?
Yeh not sure if this one I’ve seen few circulating
Avatar
so are we waiting for openrouter to add it officially, or are we waiting for them be ready?
Avatar
Avatar
so are we waiting for openrouter to add it officially, or are we waiting for them be ready?
They secured listing on Lmarena
Avatar
They don’t really need open router tho it would be good to see it on OR as my app is running via OR
Avatar
Craig Federighi 11/7/2025 2:59 PM
still fake product
Avatar
Avatar
deen
so?
BRO LM arena is the biggest
2:59 PM
Open router is a wrapper for every API basically nothing special
Avatar
Avatar
deen
so?
so they are waiting to publish lmarena results maybe
Avatar
It’s like notfiamond
Avatar
there is no mention of lmarena though
Avatar
Craig Federighi 11/7/2025 3:00 PM
Movement AI in a nutshell: release a product, claim its the best. Claim bespoke hardware which would require hundreds of millions of R&D. ZERO internal or public benchmarks.
☝️ 1
Avatar
Avatar
there is no mention of lmarena though
In their group they posted they are adding it into lmarena
Avatar
Avatar
movementlabsAI
we are not here to chase your small bucks. we are here to shape the future.
also sus
Avatar
Craig Federighi 11/7/2025 3:01 PM
the fact is that they are incapable of making up benchmark numbers themselves
Avatar
cairo com 2 a & 2 i 11/7/2025 3:01 PM
"never use em dashes" procedes to use em dashes in the instruction
Avatar
Avatar
also sus
every single message they wrote is extremely sus
☝️ 1
Avatar
Avatar
deen
every single message they wrote is extremely sus
Have u actually tried it? Or nah
3:02 PM
Avatar
Avatar
Anysogere
Have u actually tried it? Or nah
do you understand why that doesn't matter?
☝️ 3
3:02 PM
well. obviously not
Avatar
Avatar
deen
do you understand why that doesn't matter?
No 💀😭
Avatar
Craig Federighi 11/7/2025 3:03 PM
fraud
Avatar
Someone said they could be stealing from cursor new model
3:03 PM
That made sense but then cursor don’t have api or do they idk
Avatar
Avatar
Anysogere
Someone said they could be stealing from cursor new model
Do you believe anything someone says?
Avatar
Avatar
KP
Do you believe anything someone says?
Nope, I tested it myself
3:03 PM
Results shocked me
🤯 1
3:04 PM
3:04 PM
Like check this shi bro
Avatar
Avatar
Anysogere
Someone said they could be stealing from cursor new model
From testing or just basic specs comparisons one would very quickly notice that this, in fact, cannot even possibly begin to make any sense
Avatar
Craig Federighi 11/7/2025 3:04 PM
you realize that the system prompt basically proves that they don't have their own model right?
3:04 PM
they were literally prompting in SimpleBench answers.
Avatar
Avatar
Anysogere
That made sense but then cursor don’t have api or do they idk
How would this possibly make any sense?
Avatar
Avatar
Craig Federighi
you realize that the system prompt basically proves that they don't have their own model right?
I know what a system prompt is. And no I disagree lol
3:05 PM
Every ai company has a system prompt
Avatar
Craig Federighi 11/7/2025 3:05 PM
You've made no attempt to prove us that you're capable of critical thinking. So, no, I don't really care
Avatar
And none is simple bench
Avatar
Claude, Google, etc
Avatar
In fact, no serious AI ever has any benchmark in the system prompt
Avatar
So it don’t really prove anything it’s all 50/50
Avatar
Craig Federighi 11/7/2025 3:05 PM
ah yes claude and gemini prompts in simpleqa answers
3:05 PM
🤣 🤣 🤣
Avatar
Avatar
Anysogere
Every ai company has a system prompt
You are conflating "every LLM has a system prompt" with "every LLM has benchmarks in the system prompt". These two things are extremely different,
Avatar
Scam Altman is behind this
💀 1
3:07 PM
Avatar
you're a child using it for free, is that it?
3:08 PM
it's pretty good in that case
3:08 PM
wtf? (the "You started by saying "I'm here to help. What can I assist you with?") (edited)
Avatar
Avatar
deen
you're a child using it for free, is that it?
No I’m actually not, lol I have Claude pro subscription
3:09 PM
Claude is my love
☝️ 1
3:09 PM
The only reason why I like this new model is because it matches Claude
3:09 PM
That’s the only reason
Avatar
ok. well, you're not the target of the scam, so, have fun
💯 3
Avatar
He is jai user judging by his Twitter profile
3:10 PM
Which is worse
3:11 PM
And explains overjoy of another moderately capable model = non-power user
3:12 PM
i suppose this marks the end of this saga as i don't expect it to release officially on openrouter
Avatar
cairo com 2 a & 2 i 11/7/2025 3:13 PM
it's so fun how this is even being discussed, even the MPU part
3:13 PM
of course it doesn't exist
3:13 PM
it's literally Cerebras inference
Avatar
Avatar
cairo com 2 a & 2 i
it's so fun how this is even being discussed, even the MPU part
The LMArena Discord is eating this up
Avatar
CPU = Cerebras Powered Unit (edited)
👆 1
Avatar
Why does the system prompt have a crypto address for donations
💀 3
Avatar
As in, they're siding with Momentum lol
Avatar
lets see this
Avatar
Avatar
Anysogere
lets see this
cairo com 2 a & 2 i 11/7/2025 3:14 PM
dude, it's just a fine tuned model
Avatar
Avatar
Flowsen
Why does the system prompt have a crypto address for donations
Pls put coin in my tip jar so I can beat Sama ty
Avatar
cairo com 2 a & 2 i 11/7/2025 3:14 PM
open source models are also that good
Avatar
Avatar
KP
Pls put coin in my tip jar so I can beat Sama ty
You wanna put your 'tip' in a 'jar'?
Avatar
KIMI K2
Avatar
Avatar
the burden of proof is on you, not them
I have the most powerful quantum computer in the world, developed by me and me only, in my possession
3:17 PM
I'm running my custom-trained ASI on it
Avatar
cairo com 2 a & 2 i 11/7/2025 3:17 PM
ok i believe you
Avatar
Avatar
cairo com 2 a & 2 i
ok i believe you
Exactly
Avatar
Avatar
Flowsen
I have the most powerful quantum computer in the world, developed by me and me only, in my possession
I trust you, where can I pay for access? I have $3
Avatar
I keep it for me and my cat
Avatar
cairo com 2 a & 2 i 11/7/2025 3:17 PM
you have the expert role so i trust you
💯 1
Avatar
I translate the meows and humanspeak back and forth
Avatar
Avatar
Loinne
I trust you, where can I pay for access? I have $3
You can access mine if you give me bitcoin
🤔 1
Avatar
Avatar
KP
You can access mine if you give me bitcoin
I believe you
Avatar
TRUST IN YOUR EXPERTS
Avatar
Avatar
Flowsen
I have the most powerful quantum computer in the world, developed by me and me only, in my possession
it doesn't matter as long as their model, their product, is real, which doesn't seem.
3:18 PM
explained best
Avatar
Avatar
deen
TRUST IN YOUR EXPERTS
There are so many in this thread, it's like a Mixture of Experts here
😂 1
✍️ 1
Avatar
Avatar
explained best
cairo com 2 a & 2 i 11/7/2025 3:19 PM
yeah that's basically what we said on day 1
3:19 PM
lol
Avatar
Avatar
Loinne
There are so many in this thread, it's like a Mixture of Experts here
Avatar
Avatar
explained best
so its an actual model of their own hosted somewhere fast fine tuned
Avatar
Avatar
it doesn't matter as long as their model, their product, is real, which doesn't seem.
Ok I can see that they have some model running. They already proved that. Do they have some custom hardware running though? Is it really their model?
Avatar
if so that would still sell
Avatar
Avatar
Anysogere
if so that would still sell
cairo com 2 a & 2 i 11/7/2025 3:19 PM
YEAH!
Avatar
why don't they just say that if thats the case
Avatar
Avatar
Anysogere
why don't they just say that if thats the case
cairo com 2 a & 2 i 11/7/2025 3:19 PM
YES!
3:19 PM
that's the issue congrats
Avatar
ohhhh
Avatar
Avatar
Anysogere
so its an actual model of their own hosted somewhere fast fine tuned
deepseek tuned on claude/gpt-4o hosted in cerabras is my best guess at this point
Avatar
Avatar
Anysogere
why don't they just say that if thats the case
I would have no issue with a company reselling a finetuned model
Avatar
cairo com 2 a & 2 i 11/7/2025 3:20 PM
they're trying to sell themselves as the next step in AI
3:20 PM
when it's not
3:20 PM
of course that draws more attention than a fine tuned model
Avatar
Avatar
Anysogere
why don't they just say that if thats the case
Because lies sell better
Avatar
Avatar
cairo com 2 a & 2 i
of course that draws more attention than a fine tuned model
They want to get some flash attention to them
🔥 3
✍️ 2
Avatar
Avatar
Flowsen
Because lies sell better
true true
3:21 PM
i still remeber when they said grok 4 is agi
Avatar
Avatar
Flowsen
Ok I can see that they have some model running. They already proved that. Do they have some custom hardware running though? Is it really their model?
true.
Avatar
As I said before: it definitely one of the models of all time (edited)
Avatar
Avatar
Anysogere
😂
bro is trying to get scammed on discord but they don't even notice him 😭
😭 3
Avatar
i topped up api credits 👀
🧠 1
💸 1
Avatar
Avatar
Loinne
As I said before: it definitely one of the models of all time (edited)
with reflection 70b being first (edited)
Avatar
Avatar
with reflection 70b being first (edited)
nah 70 b models are retarded
Avatar
Avatar
with reflection 70b being first (edited)
i was trying to remember if that was on lmarena too
Avatar
Ok, real talk, how expensive is to fine-tune a big model like GLM 4.5?
Avatar
Avatar
Anysogere
nah 70 b models are retarded
Size is not the point of this, did you understand the reference? Do you know what Reflection 70B is?
Avatar
Avatar
Loinne
Ok, real talk, how expensive is to fine-tune a big model like GLM 4.5?
probbaly 100k?
3:24 PM
or more
Avatar
Avatar
KP
Size is not the point of this, did you understand the reference? Do you know what Reflection 70B is?
i think they're still processing what's happened here
Avatar
im asking claude hang on
Avatar
Avatar
KP
Size is not the point of this, did you understand the reference? Do you know what Reflection 70B is?
cairo com 2 a & 2 i 11/7/2025 3:25 PM
i don't recall that, but someone mentioned here they were getting flashbacks
Avatar
Cloud GPU Costs (AWS/GCP/Azure): A100 80GB: ~$30-40/hour per GPU For 8x A100s: ~$240-320/hour If training takes 2-4 weeks continuously: $40,000-215,000+
Avatar
wasn't claude the first to call sus in this thread
Avatar
cairo com 2 a & 2 i 11/7/2025 3:26 PM
yes haha
Avatar
Avatar
deen
wasn't claude the first to call sus in this thread
wdym?
Avatar
cairo com 2 a & 2 i 11/7/2025 3:26 PM
it was like NUH UH No way
👀 1
Avatar
Avatar
wdym?
cairo com 2 a & 2 i 11/7/2025 3:26 PM
i asked on perplexity with claude 4.5
👌 1
3:26 PM
about the claims, which they removed from the website later
Avatar
Avatar
wdym?
When asked about this company's specs, Claude said it's fake
👌 1
Avatar
Avatar
cairo com 2 a & 2 i
about the claims, which they removed from the website later
which date? need to check wayback machine
Avatar
Avatar
KP
When asked about this company's specs, Claude said it's fake
4.5 is self-aware and smart as fuck, i worship it like deity
💀 2
Avatar
Avatar
cairo com 2 a & 2 i
Click to see attachment 🖼️
cairo com 2 a & 2 i 11/7/2025 3:27 PM
.
Avatar
granted it also didn't realise that Cerebras was real, but...
3:27 PM
but they're definitely real.
3:27 PM
right?
3:28 PM
right.
Avatar
cairo com 2 a & 2 i 11/7/2025 3:28 PM
i used search, and it compared with cerebras and groq immediately
Avatar
cairo com 2 a & 2 i 11/7/2025 3:28 PM
i asked about Cerebras and why couldn't anyone replicate what they had yet
Avatar
cairo com 2 a & 2 i 11/7/2025 3:28 PM
Cerebras people have a lot of work before
Avatar
Avatar
cairo com 2 a & 2 i
.
blockchain infrastructure
yeah %100 fake
Avatar
Avatar
blockchain infrastructure
yeah %100 fake
cairo com 2 a & 2 i 11/7/2025 3:29 PM
read the rest, it couldnt access the website at first
3:29 PM
i had to copy paste it
3:29 PM
and it couldnt find anything about them on the web
Avatar
the blcockhain stuff is from another company with same name
☝️ 1
3:30 PM
thats why the ai's get confused lol
Avatar
cairo com 2 a & 2 i 11/7/2025 3:30 PM
that's only for the first message
Avatar
they are not agi yet
Avatar
cairo com 2 a & 2 i 11/7/2025 3:30 PM
🫩
Avatar
at this point it doesn't matter, we are kicking a corpse, i'll go make dinner and eat the rest of this thread
🤝 1
Avatar
Avatar
at this point it doesn't matter, we are kicking a corpse, i'll go make dinner and eat the rest of this thread
Don't eat this thread
😭 1
Avatar
Yummy slop
Avatar
Avatar
Anysogere
the blcockhain stuff is from another company with same name
you sure about that?
Avatar
this is all PR for them
3:31 PM
all PR is good PR😂
3:31 PM
they say
Avatar
cairo com 2 a & 2 i 11/7/2025 3:31 PM
dumb people say that yes
Avatar
hahaha
Avatar
4. Funding & Economics Zero ICO, zero presale, zero "investor" wallets. Pricing is posted in plain sight on our site, payable per token via Stripe or crypto. Donations are voluntary tips exactly like GitHub Sponsors or Ko-fi for open-source tools. And why would be offer $1/m in and $1/m out?
Avatar
cairo com 2 a & 2 i 11/7/2025 3:32 PM
good PR is always better than bad PR long term
Avatar
because people obviously aren't reading things
Avatar
open router loves crypto too
Avatar
i love my cerebras investor wallets
Avatar
doesn't make them a scam
3:33 PM
just cus something has "crypto" in it is not a scam dude
Avatar
cairo com 2 a & 2 i 11/7/2025 3:34 PM
i don't think that's the point
Avatar
you're funny
Avatar
No one is saying that this in particular makes it a scam
3:34 PM
Idk where you draw 80% of your conclusions from
Avatar
i know i'm just saying
Avatar
cairo com 2 a & 2 i 11/7/2025 3:34 PM
why are you using their tag lol 😭
Avatar
becuase only 4-5 people are attacking movement guys and then they also have a lot of support i support them
Avatar
cairo com 2 a & 2 i 11/7/2025 3:35 PM
attacking dude lmaop
Avatar
model is good and speed is good so yeh
Avatar
Good old bandwagon effect
Avatar
thats all that matters for me
Avatar
If everyone likes it, that must be true
Avatar
cairo com 2 a & 2 i 11/7/2025 3:35 PM
if so that would still sell why don't they just say that if thats the case
!!!
Avatar
Avatar
cairo com 2 a & 2 i
if so that would still sell why don't they just say that if thats the case
!!!
this yes.
Avatar
cairo com 2 a & 2 i 11/7/2025 3:35 PM
we wouldn't be discussing the legitimacy of this
3:36 PM
if they were clear
3:36 PM
but they provided no proof
Avatar
maybe the fucked up and are too deep
3:36 PM
in
3:36 PM
to go back
Avatar
cairo com 2 a & 2 i 11/7/2025 3:36 PM
the proof they provided only strenghtened my belief that its a cerebras inference
Avatar
Avatar
Anysogere
thats all that matters for me
Nice, good to know There are some people that value integrity and transparency, but if you value these things, all the power yo you I guess
Avatar
how much does it cost to host there?
3:37 PM
i wana fine tune a model now
💀 1
Avatar
Avatar
Anysogere
maybe the fucked up and are too deep
if it's actually even a finetune, then yeah they would have outlaid a stack
3:39 PM
^^ more proof towards a fine tune for sure
3:39 PM
people are switching from glm to momentum
3:39 PM
3:40 PM
they scraped data from claude api in that case
Avatar
but the average iq of glm users is increasing??
Avatar
and that's againgst clude rules, to build a competing product. i swear openai had backlash of this some time ago when their api got revoked (edited)
Avatar
Where. Benchmarks
Avatar
Simple-bench 100%
💯 2
Avatar
cairo com 2 a & 2 i 11/7/2025 3:42 PM
ok
Avatar
Oh, boy, if this goes mainstream 🍿
Avatar
Avatar
KP
Oh, boy, if this goes mainstream 🍿
its gona be real drama world wide lol
3:43 PM
3:44 PM
im just going to wait for benchamrks on lm arena until them in gona use it as much as i can to complete somes games i have pending
Avatar
What if we just make a model with long context and put answers from all benches into system prompt?
Avatar
Avatar
Anysogere
Click to see original message
we're pleased to announce that we will have have an upcoming announcement regarding LMarena very soon
Avatar
Avatar
Loinne
What if we just make a model with long context and put answers from all benches into system prompt?
we should all get together and make our own LLM
3:48 PM
call it benchmarkbuster v1
Avatar
Very_serious_person 11/7/2025 3:54 PM
Avatar
CrushedAsian255 11/7/2025 3:59 PM
Is this actually a good model or is it benchmaxxed (edited)
Avatar
I recommend giving thia thread a read lol
4:00 PM
Just use GLM or something
Avatar
he loved your humor though 😭
Avatar
Craig Federighi 11/7/2025 4:01 PM
is a scam
Avatar
☝️ this is why im scared for our future
Avatar
Let's see who laughs last when it scores 2000+ on eqbench
👍 1
Avatar
https://www.youtube.com/watch?v=qSqAxEXtZY0 i take everything back, this video about the MPU is very inspiring
4:28 PM
now these are the ai men we need
Avatar
cairo com 2 a & 2 i 11/7/2025 4:31 PM
which video
4:31 PM
oh haha
4:31 PM
the MPU yes
Avatar
https://www.youtube.com/watch?v=eIwzjvTy8KE god damn. the music on this datacenter video is JAUNTY
6:27 PM
6:27 PM
wew guys, look our model now says it's system prompt out loud now (edited)
👍 1
6:31 PM
😱 1
6:32 PM
release the benchmarks already
Avatar
cairo com 2 a & 2 i 11/7/2025 6:34 PM
so generous
Avatar
Avatar
movementlabsAI
Click to see attachment 🖼️
see, we have british man in our powerpoint presentation. doubt us? we will post the same picture. https://canary.discord.com/channels/1340554757349179412/1340554757827461211/1436122028095832167 we worked on a browser nobody heard of. (edited)
Avatar
cairo com 2 a & 2 i 11/7/2025 6:46 PM
a sheikh isn't all you need to create a industry-leading chip and technology
6:47 PM
Cerebras has a 400+ employees and launched their hardware on 2021
6:48 PM
last time it was 700b+
Avatar
Avatar
Click to see original message
thats bit more than 700B, nearly double
☝️ 2
Avatar
cairo com 2 a & 2 i 11/7/2025 6:55 PM
is it just like that?
6:55 PM
one checkpoint has X billion parameters and you just add it to the next?
6:56 PM
genuine question
Avatar
I don't think that how it works, to increase parameters you need to re-train. At least that's what I heard
Avatar
Avatar
Loinne
I don't think that how it works, to increase parameters you need to re-train. At least that's what I heard
Yes that's what I believe too
☝️ 1
6:58 PM
Nobody knows which weights are responsible for which parts of the model's intelligence
6:59 PM
This requires separate research into an already trained model
Avatar
cairo com 2 a & 2 i 11/7/2025 6:59 PM
also, parameters isn't everything, case in point Ling/Ring
Avatar
You can know if trained on MPU
Avatar
Avatar
Loinne
You can know if trained on MPU
Magic Processing Unit
Avatar
Avatar
Loinne
I don't think that how it works, to increase parameters you need to re-train. At least that's what I heard
yes you need to retrain in most cases but if its an moe you could technically finetune the router & the model and add extra experts but thats quite sketchy
Avatar
I think we have to wait for real benchmarks. Currently, it looks quite suspicious. A company that has its own chip and builds its own large LLM model doesn’t seem to have money for a proper website or to evaluate its own model. Nevertheless, I find the model quite impressive - it feels better than other open-source models. If they set the pricing at the promised level ($1 in, $1 out), it could become a very interesting alternative to other models. We’ll have to wait and see.
☝️ 1
Avatar
Avatar
movementlabsAI
fair and square.
this is hilarious lol
Avatar
mr movement is offline today, no news
2:35 PM
Avatar
Avatar
mr movement is offline today, no news
No they go to Imarena discord
2:43 PM
No on here now
Avatar
no no, even there he is offline
Avatar
Avatar
no no, even there he is offline
But he messages on imarena today but not here
Avatar
Avatar
buun
But he messages on imarena today but not here
was 11 hours ago
Avatar
Have a trust!
😭 1
Avatar
movementlabsAI 11/8/2025 3:24 PM
Quick update: We've changed direction and won't be pursuing OpenRouter integration. Thanks for the opportunity! We are in talks with some other companies that will make better use of free credits from us.
😆 5
😭 2
Avatar
cairo com 2 a & 2 i 11/8/2025 3:25 PM
oh no! we're so ungrateful to miss on this opportunity 😓
😭 3
Avatar
movementlabsAI 11/8/2025 3:26 PM
I don't understand why this thread is still open 🙂
Avatar
cairo com 2 a & 2 i 11/8/2025 3:26 PM
for future research
☝️ 1
Avatar
Avatar
movementlabsAI
I don't understand why this thread is still open 🙂
You can close it
Avatar
But free credits...
😢 1
3:28 PM
and the saga ends here, was fun while it lasted. (edited)
Avatar
Avatar
movementlabsAI
Quick update: We've changed direction and won't be pursuing OpenRouter integration. Thanks for the opportunity! We are in talks with some other companies that will make better use of free credits from us.
nooooooo
6:21 PM
why couldn't have we just WAITED for the BENCHMARKS
☝️ 1
Avatar
We are in talks with some other companies that will make better use of free credits from us
Nice PR with other companies there, implying OR won't make good use of free credits
Avatar
I am waiting for movementlabs / orbiousai collab
Avatar
We Are In Talks with numerous other online routers, arenas, and other circular websites and are awaiting a call back at any moment now
😭 1
😂 1
Avatar
BeetleBonsai798 11/8/2025 6:56 PM
+
Avatar
wait this thread was still up? lol
Avatar
Guys, we are not here to waste time with some unknown people. Whatever personal vendetta you might have will be addressed on our official channels.
this time... it's personal.
⚔️ 1
Avatar
cairo com 2 a & 2 i 11/9/2025 3:04 AM
it's really not. the only personal aspect is that i find funny how a company with this type of technology would discuss with random people on the internet instead of just launching everything at once (edited)
Avatar
I just imagine how hilarious it would be if a popular AI lab engaged like this with random potential customers
Avatar
Avatar
cairo com 2 a & 2 i
it's really not. the only personal aspect is that i find funny how a company with this type of technology would discuss with random people on the internet instead of just launching everything at once (edited)
i'm sure they put a lot of effort into setting up the inference, but it probably feels very personal because this technology mainly exists in their heads
😭 1
Avatar
Avatar
movementlabsAI
You're absolutely right ignore the haters! 😄 Yeah, we spent hundreds of millions building the world's best AI chip with zero chip-making experience. Totally reckless move. And sure, we posted some numbers that were... let's call them 'optimistic projections'... then quietly updated them when people started doing math. Shocking behavior, I know. And adding a donation button to our $8 billion product? Classic overpriced tip jar move. We're basically the Elon Musk of unnecessary funding options. But hey, at least the chip works better than your average potato, and we're actually delivering real inference speeds that beat the competition. So... worth it? 😏
open source wen?
Avatar
we got this thread before gta6
Avatar
wow they added over 60B parameters in one day.. they must have had a lot of momentum to be able to go that fast
Avatar
Avatar
newman
wow they added over 60B parameters in one day.. they must have had a lot of momentum to be able to go that fast
Solenopsisbot 11/9/2025 7:48 AM
second screenshot sketchy as hell
7:49 AM
why would quant change parameters?
Avatar
Avatar
Solenopsisbot
second screenshot sketchy as hell
i think its fine, just a shortened version of "its above 700B parameters, and its at FP16", not "its above 700B parameters when its at FP16"
7:52 AM
a little ambiguously worded but there is a reasonable plausible interpretation
Avatar
Avatar
newman
wow they added over 60B parameters in one day.. they must have had a lot of momentum to be able to go that fast
That's so funny, asking them got them very confused
3:38 PM
3:39 PM
Avatar
It does not sound like qwen/glm merge
Avatar
movementlabsAI 11/10/2025 3:04 AM
preview of checkpoint 4
3:08 AM
3:09 AM
3:10 AM
You have expressed interest in leaving this group, said you were going to give up on a partnership with OR and said you are in talks with companies that will make "better" use of free credits
Avatar
movementlabsAI 11/10/2025 3:10 AM
well i deleted it a few times and you guys bring it back, so why not post in here from time to time?
3:10 AM
free PR
Avatar
Yeah, I suppose you can keep advertising changes for PR here even though you have no interest in pursuing any OpenRouter integration, if you wish
3:14 AM
Good luck with your quest in getting "free PR", and to whoever sends more messages here that help you achieve your goal
Avatar
i appreciate the updates
Avatar
Avatar
movementlabsAI
preview of checkpoint 4
Movement World! can you make it multiplayer so we can all join?
Avatar
forgortochngeit 11/10/2025 3:30 AM
With blackjack?
Avatar
yes, and a science-based dragon mmo
Avatar
Avatar
movementlabsAI
well i deleted it a few times and you guys bring it back, so why not post in here from time to time?
we are bringing it back?
Avatar
Avatar
movementlabsAI
well i deleted it a few times and you guys bring it back, so why not post in here from time to time?
your company was allegedly exposed as a scam, everyone lost their interest and forgotten afterward. any pr is a bad pr for you and you tarnished the trust of your 'potential' customers. wherever you go, people will remember of your shady company, and no one reputable will partner with your company. (edited)
Avatar
cairo com 2 a & 2 i 11/10/2025 1:17 PM
"make it better" 🐲
1:18 PM
procedes to make the most sloppiest AI coded welcome screen ever
1:18 PM
is that comic sans
Avatar
Lol, movementlabs left
Avatar
we did it joe
Avatar
Avatar
cairo com 2 a & 2 i
"make it better" 🐲
He was referring to checkpoint 4, not the current checkpoint 3, but I don't know if that's true. (edited)
Avatar
Avatar
buun
He was referring to checkpoint 4, not the current checkpoint 3, but I don't know if that's true. (edited)
cairo com 2 a & 2 i 11/10/2025 1:55 PM
that's from his video
2:10 PM
OR ppl are saved from the scam, we won.
2:11 PM
2:12 PM
now he is at yupp when OR and lmarena failed them, hilarious, lol (edited)
Avatar
cairo com 2 a & 2 i 11/10/2025 2:13 PM
can you send me the link lol
Avatar
Avatar
cairo com 2 a & 2 i
can you send me the link lol
sure, join using this: yuppai (edited)
Avatar
cairo com 2 a & 2 i 11/10/2025 2:14 PM
thanks
❤️ 1
Avatar
his whole history (edited)
Avatar
Avatar
movementlabsAI
we are not here to chase your small bucks. we are here to shape the future.
meanwhile him days ago: (edited)
Avatar
cairo com 2 a & 2 i 11/10/2025 2:18 PM
the contrast between the amateurism of this account and proofs, and the claims they made
2:18 PM
is comedy
Avatar
like, peak content. so much for the reputable company that they are, who need others to do the benchmarking for them.
Avatar
Nooo my favourite lab left
3:32 PM
Cant have nice things 🙄
Avatar
forgortochngeit 11/10/2025 3:34 PM
Did they leave any chips
Avatar
Avatar
assuming it IS model agents (though this doesn't prove anything), it makes sense why they didn't released the benchmark, but their speed is fast so... i lean on deepseek tuned on gpt-4o/claude theory as it makes more sense. (edited)
They've literally already said what it is and its not that https://discord.com/channels/1091220969173028894/1434917422686801980/1435067699251249193
Avatar
Avatar
Kami
They've literally already said what it is and its not that https://discord.com/channels/1091220969173028894/1434917422686801980/1435067699251249193
architecturally not compatible, obviously he is bullshitting.
Avatar
Avatar
architecturally not compatible, obviously he is bullshitting.
LMAO i didnt know that
👌 1
4:21 PM
milkLeAwkward
Avatar
translation: we asked claude how to fine tune a model and serve it on cerebras
Avatar
Hey everyone, I'm late to the party. What's the TL;DR on the Momentum drama?
12:03 AM
Seems like they demoted everyone in their server now, their mods were fierce defenders
12:03 AM
Wonder if these people will keep defending now
Avatar
Momentum Stalking ^
Avatar
cairo com 2 a & 2 i 11/11/2025 12:35 AM
🍵
Avatar
Avatar
GMYGOM
Hey everyone, I'm late to the party. What's the TL;DR on the Momentum drama?
  • fraud amateurish company with false advertising, the more he talked the more he slip
  • ppl bully the arab scammer and he acts smug about it
  • man got tired and left and goes another places to scam people
  • everyone still laughing
  • the end if have free time, worth a read, a case study even.
(edited)
THE END?