正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Local Laya beat OpenAI's decision API 🕷 Four decision models raced down a web page ripping out links that name a spider and Laya running locally on an M5 Max took 34 ms a link so it checked 3x more than the cloud models Run decision models locally ->

atomic.chat

18,564 subscribers

216,880 次观看 • 12 天前 •via X (Twitter)

新闻政治科学技术

Anya Rossi• Live Now

Private livecam show

26 条评论

AVB 的头像

AVB12 天前

@OpenAI The first spider for super excited to see its own name and smashed through it lol

Jigs 的头像

Jigs12 天前

@OpenAI Laya is actually good for its size

Omar 的头像

Omar12 天前

@OpenAI Local gutsy outperforms local laya, and is also free, open source, and just 500MB for CPU inference

Secure the AI 🤖 的头像

Secure the AI 🤖12 天前

@OpenAI

Liam Vinberg 的头像

Liam Vinberg12 天前

@OpenAI depends on the task

Thusatharan 的头像

Thusatharan12 天前

@OpenAI no network round trip per call, of course local takes it

Mateleo 的头像

Mateleo12 天前

@OpenAI kudo for the viz, wow

Mattia 的头像

Mattia12 天前

@OpenAI I‘m using laya for my Hermes Agent and it’s perfect

Vikas 的头像

Vikas12 天前

@OpenAI Decider-4B is better model than laya. Try using it.

Schwarzer Ritter 的头像

Schwarzer Ritter12 天前

@OpenAI small = fast

Roko ʕ •ᴥ•ʔっ🪄✨🐍 的头像

Roko ʕ •ᴥ•ʔっ🪄✨🐍12 天前

@OpenAI I wonder how DM-JEPA would do here

Jaid 的头像

Jaid12 天前

@OpenAI S-tier eval visualization 💎

Martin Ronfort 的头像

Martin Ronfort12 天前

@OpenAI Laya running locally is actually proving decision models can outpace cloud APIs on both speed and cost. We went deeper on this shift in how AI judgment is changing here:

promovierdichinsknie 的头像

promovierdichinsknie12 天前

@OpenAI i agree in my test laya also outperformed clef flash an CLM

Yoeven 的头像

Yoeven12 天前

@OpenAI You haven’t put lev to the test

Georgy GLuhov 的头像

Georgy GLuhov12 天前

@OpenAI I’ve achieved four hours of successful autonomous operation for Laya in my RimWorld colony without any input from me; all experiment logs are recorded in the repository, and I’m also planning to host a stream!

Wolkyen 的头像

Wolkyen12 天前

@OpenAI What about Mercury decide? Its the best!

Eugene Meidinger 的头像

Eugene Meidinger12 天前

@OpenAI Nightmare fuel

henrik the vibe 的头像

henrik the vibe12 天前

@OpenAI You've got me on my knees, Layla Begging darling please, Layla Darling, won't you ease my worried mind?

Brjan | AI Builder 的头像

Brjan | AI Builder12 天前

@OpenAI if local models can process links 3x faster, cloud solutions need to catch up

Toasterson @Toasterson@chaos.social 的头像

Toasterson @[email protected]12 天前

@OpenAI How did you do the animations?

Don Rai 的头像

Don Rai12 天前

@OpenAI But I tested yesterday and Laya didn't work so well 🫠

TslaMan 的头像

TslaMan12 天前

@OpenAI Nice marketing so cute

Zuhayr Khan 的头像

Zuhayr Khan12 天前

@OpenAI Did all four models pick the same links, or were there differences in what they missed?

Mark Kashef 的头像

Mark Kashef12 天前

@OpenAI I doubt it beats it on speed, Decisions API is pretty fast

nicholas gogerty 的头像

nicholas gogerty12 天前

@OpenAI how about something that scales as it fails. JEV fail-up across scales

相关视频

Local Laya blitzed other models on RTX 4090 ⚡️ Four local decision models raced down Wikipedia picking out every word that names a centipede and Laya took 3.9 ms a word so it ended up being the fastest overall Run decision models locally ->

Local Laya blitzed other models on RTX 4090 ⚡️ Four local decision models raced down Wikipedia picking out every word that names a centipede and Laya took 3.9 ms a word so it ended up being the fastest overall Run decision models locally ->

atomic.chat

26,773 次观看 • 11 天前

Local Laya moggs Jev at Grok 4.7-built Tetris 🧩 An open-weights System One model called Laya, beat cloud-based Jev at playing Tetris by making decisions 11 times faster, running locally on a 16GB MacBook Air! Run AI models locally ->

Local Laya moggs Jev at Grok 4.7-built Tetris 🧩 An open-weights System One model called Laya, beat cloud-based Jev at playing Tetris by making decisions 11 times faster, running locally on a 16GB MacBook Air! Run AI models locally ->

atomic.chat

1,098,057 次观看 • 28 天前

Laya plays Plants vs Zombies on CPU 🧟 We patched the llama.cpp engine to run Laya natively so it works on a regular CPU with no GPU and here a fine-tuned Laya makes every move until it beats the Plants vs Zombies pool level Run decision models locally->

Laya plays Plants vs Zombies on CPU 🧟 We patched the llama.cpp engine to run Laya natively so it works on a regular CPU with no GPU and here a fine-tuned Laya makes every move until it beats the Plants vs Zombies pool level Run decision models locally->

atomic.chat

14,071 次观看 • 17 天前

The open-source local version of Laya crushes Jev in response speed! On the left is Laya, a local 421M open-source decision model. On the right is Jev 1.13.0, running through a cloud API. Same Snake game, same typed decisions, and the same 30-second free run: - Laya: Score 46, Length 52, 86.5 decisions/sec - Jev: Score 1, Length 7, 3.2 decisions/sec - In a real test on an M5 Pro, median latency was 15.3ms locally versus 298.1ms in the cloud, so nearly 20x faster. Laya uses around 1GB of memory, responds in milliseconds, and doesn’t depend on a network connection. While Jev waits for a cloud round-trip at every step, Laya makes each decision instantly on-device.

The open-source local version of Laya crushes Jev in response speed! On the left is Laya, a local 421M open-source decision model. On the right is Jev 1.13.0, running through a cloud API. Same Snake game, same typed decisions, and the same 30-second free run: - Laya: Score 46, Length 52, 86.5 decisions/sec - Jev: Score 1, Length 7, 3.2 decisions/sec - In a real test on an M5 Pro, median latency was 15.3ms locally versus 298.1ms in the cloud, so nearly 20x faster. Laya uses around 1GB of memory, responds in milliseconds, and doesn’t depend on a network connection. While Jev waits for a cloud round-trip at every step, Laya makes each decision instantly on-device.

Simplifying AI

24,406 次观看 • 28 天前

Local Liquid d1 3B crushed OpenAI Decision at Snake 🐍 Liquid AI d1 on a single RTX 3090 decides in 8 ms so its snake made 82 moves a second and ate 86 apples while OpenAI's API took 345 ms a call and ate 3 Run decision models ->

Local Liquid d1 3B crushed OpenAI Decision at Snake 🐍 Liquid AI d1 on a single RTX 3090 decides in 8 ms so its snake made 82 moves a second and ate 86 apples while OpenAI's API took 345 ms a call and ate 3 Run decision models ->

atomic.chat

194,199 次观看 • 12 天前

OpenAI Decision crushed by Microsoft-Decision-1 at Space Invaders 👾 Decision-1 decided 3.5x faster so it scored 3,320 and cleared 4 waves while OpenAI's decision API scored 1,300 taking over 300ms per decision on average Run decision models ->

OpenAI Decision crushed by Microsoft-Decision-1 at Space Invaders 👾 Decision-1 decided 3.5x faster so it scored 3,320 and cleared 4 waves while OpenAI's decision API scored 1,300 taking over 300ms per decision on average Run decision models ->

atomic.chat

91,043 次观看 • 9 天前

decision models + local ai = ❤️ tried Laya (421M) on llama.cpp in a little apple-picking benchmark, it's almost as good as Qwen 4B at 12x the speed.

decision models + local ai = ❤️ tried Laya (421M) on llama.cpp in a little apple-picking benchmark, it's almost as good as Qwen 4B at 12x the speed.

Victor M

10,434 次观看 • 13 天前

Next up on First Pass: a conversation with Sam Crowder from LangChain on decision models. Jev from TypeSafe AI took the world by storm. Then OpenAI and Databricks released producs (Decisions API and ai_decide function) So what are Decision Models? We dig in!

Next up on First Pass: a conversation with Sam Crowder from LangChain on decision models. Jev from TypeSafe AI took the world by storm. Then OpenAI and Databricks released producs (Decisions API and ai_decide function) So what are Decision Models? We dig in!

Jamin Ball

12,147 次观看 • 11 天前

I got a fully maxed out MacBook, mostly so I can run local models fast. And omg here is mixtral running on ollama... almost cannot believe how fast it is. This is with no internet!! A model that beats GPT-3.5 running locally! What!

I got a fully maxed out MacBook, mostly so I can run local models fast. And omg here is mixtral running on ollama... almost cannot believe how fast it is. This is with no internet!! A model that beats GPT-3.5 running locally! What!

Charlie Holtz

196,086 次观看 • 2 年前

For anyone wandering what does it mean to run ds4-agent locally on an M5 Max using DeepSeek V4 Flash q2-imatrix gguf model. Here's a video of ds4 updating itself, adding a way to leverage HF_HOME for gguf models. Future of Local AI is bright!

For anyone wandering what does it mean to run ds4-agent locally on an M5 Max using DeepSeek V4 Flash q2-imatrix gguf model. Here's a video of ds4 updating itself, adding a way to leverage HF_HOME for gguf models. Future of Local AI is bright!

Ivan Fioravanti ᯅ

30,452 次观看 • 4 个月前

Behold... GPT-OSS (20B) running 100% locally in your browser on WebGPU. This shouldn't be possible — but with Transformers.js v4 and ONNX Runtime Web, it is! A new class of AI apps is emerging. Zero-install, infinite distribution. Simply visit a website and run models locally.

Behold... GPT-OSS (20B) running 100% locally in your browser on WebGPU. This shouldn't be possible — but with Transformers.js v4 and ONNX Runtime Web, it is! A new class of AI apps is emerging. Zero-install, infinite distribution. Simply visit a website and run models locally.

Xenova

311,512 次观看 • 8 个月前

Introducing THIS / THAT Model 1.0. Trained by FLock. Not another chatbot. A best-in-class open-source model built to decide. 94.1% success rate on our 68-question decision test, outperforming GPT-5.6, Deepseek-4.1-flash, Kimi K3, Jev, and more. Just 1.9B parameters. Runs locally. Measured decision latency of ~31 ms on an RTX 5080 laptop GPU. We’re putting decentralized AI to work: giving developers and AI agents decision-making capabilities they can run on their own hardware. Your model. Your machine. Not every decision needs the cloud.

Introducing THIS / THAT Model 1.0. Trained by FLock. Not another chatbot. A best-in-class open-source model built to decide. 94.1% success rate on our 68-question decision test, outperforming GPT-5.6, Deepseek-4.1-flash, Kimi K3, Jev, and more. Just 1.9B parameters. Runs locally. Measured decision latency of ~31 ms on an RTX 5080 laptop GPU. We’re putting decentralized AI to work: giving developers and AI agents decision-making capabilities they can run on their own hardware. Your model. Your machine. Not every decision needs the cloud.

FLock.io

12,046 次观看 • 28 天前

It's been an amazing week for open models running on NVIDIA Local AI systems! 👀 Here's a few you’ll want to know about, all ready to run locally on NVIDIA RTX, DGX Spark, DGX Station, and Jetson.

It's been an amazing week for open models running on NVIDIA Local AI systems! 👀 Here's a few you’ll want to know about, all ready to run locally on NVIDIA RTX, DGX Spark, DGX Station, and Jetson.

NVIDIA RTX Spark

60,797 次观看 • 2 个月前

Running 1-bit Qwen 3.8 Flash Next (79GB) on a MacBook Pro M5 Max 64GB at 30 tok/s 🤯 It ran a 8-minute agent loop with 6 web searches, 3 Python runs and 5 sourced tables. Run local models via Atomic Chat! GGUF:

Running 1-bit Qwen 3.8 Flash Next (79GB) on a MacBook Pro M5 Max 64GB at 30 tok/s 🤯 It ran a 8-minute agent loop with 6 web searches, 3 Python runs and 5 sourced tables. Run local models via Atomic Chat! GGUF:

atomic.chat

103,645 次观看 • 1 个月前

Hey Theo - t3.gg my Tesla has a local model running it. It works without being connected at all. And doesn’t kill me. Just got home. Alibaba Cloud researchers told me this week at ACL that the real action in the labs is building smaller models. So the open source & “run locally” bulls will have their day.

Hey Theo - t3.gg my Tesla has a local model running it. It works without being connected at all. And doesn’t kill me. Just got home. Alibaba Cloud researchers told me this week at ACL that the real action in the labs is building smaller models. So the open source & “run locally” bulls will have their day.

Robert Scoble

18,446 次观看 • 3 个月前

Genspark has released a new AI browser that gives you access to: - 169 open source models that run LOCALLY - Automation of browsing and workflows - Agents on any page And much more. Available on Mac and Windows (use cases below)

Genspark has released a new AI browser that gives you access to: - 169 open source models that run LOCALLY - Automation of browsing and workflows - Agents on any page And much more. Available on Mac and Windows (use cases below)

Paul Couvert

80,955 次观看 • 1 年前

We turned Qwen3.8-27B into a multimodal decision model. It beat Pokémon FireRed’s elite four and champion with sub-100 ms decisions from live game state. With SGLang’s native /v1/decisions, you can now turn LLMs and VLMs into classification and scoring models. We also added /v1/systemone so Jev-like open models can work with the TypeSafe SDK.

We turned Qwen3.8-27B into a multimodal decision model. It beat Pokémon FireRed’s elite four and champion with sub-100 ms decisions from live game state. With SGLang’s native /v1/decisions, you can now turn LLMs and VLMs into classification and scoring models. We also added /v1/systemone so Jev-like open models can work with the TypeSafe SDK.

SGLang

252,578 次观看 • 19 天前

ParoQuant just got a big upgrade 🚀 ✅ Supports the new Qwen3.5 models ⚡ Now runs on MLX (fast local inference on Apple Silicon) 🧠 Preserves reasoning quality with 4-bit quantization We also built an agent demo running locally on my 4-year-old M2 Max. Can't wait to upgrade to an M5 Max and see what kind of magic we can do. ✨

ParoQuant just got a big upgrade 🚀 ✅ Supports the new Qwen3.5 models ⚡ Now runs on MLX (fast local inference on Apple Silicon) 🧠 Preserves reasoning quality with 4-bit quantization We also built an agent demo running locally on my 4-year-old M2 Max. Can't wait to upgrade to an M5 Max and see what kind of magic we can do. ✨

Zhijian Liu

49,826 次观看 • 7 个月前

You can run local AI models up to 120B without a $10,000 Mac Studio This phone-sized device is your own server for open-source models. 100% local and private. And you can use it to: - Power an agent like OpenClaw 24/7 - Completely replace a chatbot - Literally anything that requires an API It’s called Tiiny, and, once again, you can run the latest open-source models on it. Link below

You can run local AI models up to 120B without a $10,000 Mac Studio This phone-sized device is your own server for open-source models. 100% local and private. And you can use it to: - Power an agent like OpenClaw 24/7 - Completely replace a chatbot - Literally anything that requires an API It’s called Tiiny, and, once again, you can run the latest open-source models on it. Link below

Paul Couvert

41,369 次观看 • 7 个月前

We just launched Telnyx Decision Models in beta. AI agents make small judgment calls on every run. Is this caller asking for a refund? Which team owns this ticket? Is this incident urgent enough to page someone? Decision Models answers them in one API call. Send the context and up to 64 typed questions, and get typed answers back: a choice, a yes/no score, or a rating on a scale. Your code branches on the answer. There's no free text to parse. We are offering 2 models, chosen per request: → Decision Flash, the default, for high-volume routing and triage at the lowest cost and latency. $0.035 per 1M input tokens. → Decision Pro, for decisions that depend on long context, like a full support history or a long call transcript. $0.20 per 1M input tokens. Output tokens are free on both. The API is TypeSafe-compatible, so teams already using the TypeSafe SDK can point it at Telnyx with a new base URL and model name. It runs on the same API key as Telnyx voice, speech-to-text, and inference, so you don’t need to join any waitlist or go through extra setup to access it. Try it today:

We just launched Telnyx Decision Models in beta. AI agents make small judgment calls on every run. Is this caller asking for a refund? Which team owns this ticket? Is this incident urgent enough to page someone? Decision Models answers them in one API call. Send the context and up to 64 typed questions, and get typed answers back: a choice, a yes/no score, or a rating on a scale. Your code branches on the answer. There's no free text to parse. We are offering 2 models, chosen per request: → Decision Flash, the default, for high-volume routing and triage at the lowest cost and latency. $0.035 per 1M input tokens. → Decision Pro, for decisions that depend on long context, like a full support history or a long call transcript. $0.20 per 1M input tokens. Output tokens are free on both. The API is TypeSafe-compatible, so teams already using the TypeSafe SDK can point it at Telnyx with a new base URL and model name. It runs on the same API key as Telnyx voice, speech-to-text, and inference, so you don’t need to join any waitlist or go through extra setup to access it. Try it today:

Telnyx

866,183 次观看 • 25 天前