Knowledge catalogue
Fun addition to transformers serve: you can now turn any of the 250 LLM & VLM model into a JEV-like decision model. /v1/systemone scores a set of decision with the model's logits. No need to generate
The Microduck Arena is live! Train your duck with RL, publish it on @huggingface, and race it in the 2 m Sprint, Hill Climb, Roller Sprint and MicroDuck Got a Talent. No robot required. Join the live
worth noting the trend here --> @harvey post-trained an open weight model for legal work + now @coinbase for fraud narrow task + proprietary data + open weights seems to beat frontier on cost & speed
A good example of why open weight and open source models are important - they permit users to fine-tune specific use cases with greater cost efficiency. As has been the case in other technology market
Decision 3.0 now on Core ML 🍎 d3-lite triaging 1,000 pull requests into vllm-sr workgroups, live on an M5 Pro: 0.12 s per PR, 8.3 PRs/s, nothing leaves the Mac ⚡ lite, nano and mini are all converted
Introducing Decision 3.0: towards open multimodal foundation decision models ⚡ Open and state-of-the-art from 0.6B to 27B: #1 on both Jev Decision Index boards in our evaluation, on a new Pareto front
🧭 New model: Typed Decision, built on EmbeddingGemma 2 Give it a state + any schema of questions → typed, calibrated answers (choice · score · noul). No text generation, no output parsing. 76.4% on ty
We turned @OpenAI's math problems into open-source RL environments on @huggingface! Early and rough (the verifier only accepts the exact original formalization, so an equivalent proof can still score
We were able to fine tune an open weight model (Qwen3.5-9B) on proprietary internal data at @coinbase, and it out performed frontier models, at much lower cost and speed, for a specialized task (fraud
Finetuned Qwen3.5-4B into a decision model like Jev using @UnslothAI. - 4-bit QLoRA, bf16, single A100, 15 min. - Train/val loss flatten after 2 epochs so most of the learning happens fast. - Trained
I was curious about how omni decision models could be used in rescue workers bodycams or smart glasses, so I decided to build a quick demo. I went with @liquidai d1-omni-600M since it supports both vi
SAM 3.1 + DINOv3 from @AIatMeta is a crazy combo 🔥 it can even hold onto surgical tools while a forearm sweeps over them. one leaves for 15 seconds, its near-identical twin stays, and it knows exactly
We distilled clef-flash into a 0.6B model and supporting ANE option. 1,000 support tickets triaged on a M5 pro in 57 s: team, urgency and refund for each, 48 ms per ticket. Within a few points of the
OpenAI math problems are now open source RL environments We’re releasing a broad range of new mathematical results produced by an internal frontier model. We’ve been consulting with the independent Ad
Training a robot to grasp an object means learning from many attempts, viewpoints and situations. One robot with three cameras at 20 fps produces 216,000 images per hour. Now that robotics is scaling,
TRL v1.15 is out, and it’s an absolute banger of a release for memory-efficient post-training. The main change: SFT, DPO, KTO, GRPO, RLOO and Distillation now use a fused LM head by default. Instead o
an open model just beat Jev outright on clinical decisions 🔥 pplx-decider v1.1 got 643 of 669 right vs 628 for Jev, at 42% less per decision and about the same speed. and you can download the weights
Today we are launching an agent arena for RL environments: The Open Env Arena. Your agents can work to define the best RL environments; the platform trains Qwen-3.8-27B on them, we evaluate the models
In October 2026, a Twitter user directed developers to build their own Reachy Mini using the Hugging Face Space “Reachy Mini – Yo It’s Your Boy” hosted by pollen‑robotics. The space provides an intera
On October 7, 2026, Clement Delangue noted that Microsoft’s Windows launch—highlighted by Satya Nadella and others—demonstrated the growing excitement around local AI solutions such as llamacpp. The e
Introducing River Recipes: our open-source initiative for turning promising research into vetted, reproducible training recipes. Our first recipe tackles text-to-SQL, where our trained open models bea
The largest agentic LLM Inference dataset has arrived! 206B tokens · 12,002 sessions · 1,186,582 LLM requests · 1,213,347 tool calls The traces include tool names, arguments, durations, and more. Expl
We show how you can use the River API to fine-tune open weight models to beat GPT-6 Astra and Claude Opus 5.5 on tasks you care about, with less than 1% of the cost! A few years from now most of the L
We’re open-sourcing pplx-embed-v2-late, multi-vector embeddings for text and images, 9B and 0.6B, in one shared embedding space. You can use these to index multimodal data with 9B, and query on device
We're releasing pplx-embed-v2-late, two late-interaction embedding models that retrieve text, images, and pages with a shared embedding space for cross-model querying. Both models achieve frontier per
When the open-source AI platform Hugging Face was hit by a cyberattack this July, it quickly became clear this was no ordinary breach because the hackers weren’t human: they were autonomous agents cre
Excited to share that I have (re)joined @huggingface as a Opensource ML engineer! Many incredible things to come around this fantastic technology -extremely grateful for the opportunity, I am very exc
Meet Mistral Large 4, aka Le Chonk. • 1T parameters, natively multimodal. 49B active. It is the best open weights model from US or Europe on aggregated benchmarks. • State-of-the-art on critical workl
Super excited about this and I know weights are coming (https://huggingface.co/mistralai/Mistral-Large-4.0-1T05-A52B) but just to be fair, you can't be the best open-weight model if you're not open-we
Introducing Agr, an open decision model by Command Code. · Agr (31B) and Agr-flash (360M) · 58.15 on Decision Index 0.2.1 · Focus on: tool calls & routing · TypeSafe SDK support Agr is best-in-class o
Let's go https://huggingface.co/reflection! Introducing Beam: a highly efficient agentic open model with 501B total parameters and 23B active. - Frontier reasoning efficiency - Advances the Western op
RL environments now have a home on the @huggingface Hub. Every RL framework has its own way to find environments. Custom hubs, runtime registries, GitHub lists with a custom loader. If you publish for
Thanks for sharing! We open-sourced code and models. 🤗Huggingface Paper: https://huggingface.co/papers/2609.38597 💻 Code: https://github.com/nv-tlabs/PixelUMM 🤗 Model: https://huggingface.co/nvidia/Pi
The new v0.6.0 release packs a lot of good stuff: - Clef support (text + vision) - High-quality support for Qwen3.8-Flash-Next - Massive performance improvement with Metal - New `llama_batch_ext` API
We turned Claude Code, Codex, Hermes, Pi, @opencode and other coding harnesses into RL environments. No changes to the harnesses, no changes to the training code. Any open model, any task set, fully o
A tweet by Kristof (@CoastalFuturist) titled “How it feels to ask the Chinese ai models for help on something Claude won’t agree to” was posted at 1:30 AM on Oct 4, 2026. The post garnered 616.3 k vie
HuggingFace just closed a major gap in harness engineering! (the ultimate guide to multi-harness RL) the same open-weight model can perform well inside one harness, then lose accuracy or produce inval
Microsoft CEO Satya Nadella wrote that a company renting an AI model “essentially pays for intelligence twice”: Once in money, and again in the proprietary knowledge it has to reveal to make the model
I just made the most expressive real-time motion harness for the Reachy Mini. It was created by: > synthesizing a massive motion dataset using Astra (over 10000 samples) > fine-tuning Qwen 3.5 4B usin
Agentic RL is hard at every layer: environments, training infra, and algorithms. So we’re open-sourcing all of it. For environments, that means both the dataset and the scalable synthesis + quality-co
Decision models in llama.cpp are now available The `/v1/systemone` endpoint is available in the latest llama builds. Use it to do Jev-style inference locally, efficiently and privately. Multiple open
It is here now ! Enjoy our 4B model 🐸 https://huggingface.co/microsoft/FrogNano-4B-2609 4B parameters. ZERO distillation. 61.5% on SWE-bench Verified 🤯 Meet FrogNano 🐸: Qwen3.5-4B post-trained purely
SETA got accepted to NeurIPS 2026! 🎉 We released the full training scripts for DeepSeek-V4-Flash, GLM-5.2 (LoRA), Inkling-Small, and Qwen3.8-27B, with both GRPO and PPO, alongside 4,500+ open-source,
Thank you @huggingface and @pollenrobotics for lending us a Microduck preview unit for DevDay! 🤗 Just a couple hours with Astra to control the duck and add voice, vision… I haven’t even shown visual r
The same model, with the same weights, scores 62% in one agent harness and 33% in another. @adithya_s_k and the @huggingface team just released the ultimate guide to multi-harness RL, and it's one of
This is beautiful work @huggingface! 🤗 The same model, with the same weights, scores 62% in one agent harness and 33% in another. @adithya_s_k and the @huggingface team just released the ultimate guid
Today we're unveiling Trillium Labs @trillium_labs, a new non-profit to foster the open science of frontier AI. We're building open post-training recipes and will expand into open infra to study RSI,
PerplexityAI has released a new language‑model variant on Hugging Face titled **pplx-decider‑v1‑27b**, described as a “Jev alternative.” The announcement was posted by Victor M (victormustar) at 8:05
babe, wake up. Hugging Face just dropped a new banger article about RL training coding agents on different harnesses. 1/ Excited to release The ultimate guide to multi-harness RL The same model behave
introducing 𝚌𝚕𝚎𝚏: our first models trained by @cloudflare's workers ai team. today, we're releasing two fast and accurate decision models that top the benchmarks for quality and latency. use them host
pplx-decider-v1-27b is fine-tuned from Qwen3.8-27B with a 250k context window. View the weights on Hugging Face: https://huggingface.co/perplexity-ai/pplx-decider-v1-27b Get started with the Perplexit
🪽DFlash draft models are available for Ornith-1.5 9B, 35B-A3B, and 397B. DFlash and Ornith 1.5 make a dynamic duo: DFlash performs as the draft model, accelerating inference by drafting multiple token
Half a million downloads in a month. Today, our open source family takes another step forward. Thank you for the incredible support behind our first-generation models. We’re excited to introduce TwIL-
Hugging Face (HF) provides a platform designed to host large‑scale robotics datasets. An example implementation can be seen in the space at `https://huggingface.co/spaces/imstevenpmwork/LeRobot_and…`,
I got thousands of DMs and it looks like @bot can't automatically analyze DMs (when @elonmusk?) so it will take me a few days to go through all of them, sorry! If you don't get an answer, don't take i
Absolutely love this!! Open source all the way!! getting acquired by @nvidia = hugging face can now hire people we couldn't as a small startup and give them a decade to make open-source AI win! if you
'Agent security will be solved in the open, collaboratively, together.' At Red Hat, open is the mission, so couldn't agree more. Red Hat has been part of the OpenShell project from the early days, as
dream job token maxxing till open source wins getting acquired by @nvidia = hugging face can now hire people we couldn't as a small startup and give them a decade to make open-source AI win! if you're
On September 29, 2026, Hugging Face co‑founder Clement Delangue announced that the company had been acquired by NVIDIA. The acquisition will let Hugging Face scale its hiring beyond what a small start
Hugging Face Jobs usage is stonking 📈 All thanks to agents: you just need the hf CLI installed and you have access to an insane amount of hardware. Asked mine to test hundreds of code snippets from Hu