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, reward-hacking, multi-agent systems, and whatever comes next. We're built around the theory of change that you need more eyes to solve hard technical problems. We have faith in the scientific methods and communities that humanity has built, and worry that AI is becoming too closed to utilize them. Trilliums are wildflowers that bloom briefly in the spring, before the forest canopies fill out. Though they are small, they lay the foundation for the cycles of growth and nourishment through the rest of the year. At Trillium Labs, the recipes will be the slow nutrients for the seasons and the model releases will be the blooms. Building an institution dedicated to this is needed because, much as nature’s trilliums are slow to expand and grow, the open-ecosystem needs time and dedicated resources to catch up. I co-founded with with a long-time friend and collaborator Tom Zick (@thesezickbeats). We're hiring (full time + student collabs/interns), we're fundraising, and we're looking for compute. Please get in touch if you're interested in helping out. Offices based in the Bay Area and Cambridge MA, remote okay. I’m in the Bay Area until for The Curve and COLM to connect with people who are interested. We’re thankful to have initial support from Halcyon Futures and Schmidt Sciences with more funding en route to enable our ambitions of scaling. Our advisors @Thom_Wolf, @HannaHajishirzi, @gneubig and @ctnzr have been instrumental to building the ecosystem that exists today, and I’m stoked to get to keep working with them.

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 the most practical RL write-ups this year, and everything open! The trick is simple. Don't touch the harness. Point it at a proxy instead of the model. The proxy speaks all four API formats coding agents use (OpenAI Chat Completions, OpenAI Responses, Anthropic Messages, Gemini). It records the exact token ids and logprobs vLLM sampled, and you train on that. You don't change a single line of Claude Code, Codex or OpenCode. Results: 🔹 Trained across 4 harnesses at once, LFM2.5-2.6B by @liquidai went from 42% to 54% 🔹 31% fewer tool calls, thanks to a small bonus for solving tasks in fewer steps 🔹 Training in OpenCode alone took OpenCode from 34% to 58%, but the multi-harness model improved everywhere They also tried the shortcut everyone reaches for: fine-tune on 3,189 successful rollouts from Qwen3.8-27B. Imitation plateaued at 47.5%, below both RL runs. Copying a bigger model doesn't get you there. Practice does. The best part is that everything is open: the capture proxy in OpenEnv, the trainer in TRL, the tasks, the SFT data, the training code and all seven trained models. Agents will run in dozens of harnesses. Now open models can be trained for each of them, by anyone. Read it here 👇 huggingface.co/spaces/FineEn…

babe, wake up. Hugging Face just dropped a new banger article about RL training coding agents on different harnesses.

New in @huggingface Jobs docs: a Train Models page. A quick overview for you or your agent, covering Transformers, TRL, @UnslothAI, @axolotl_ai and Diffusers. Each example runs in minutes on a single A10G for around $0.10. huggingface.co/docs/hub/jobs…

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 models are supported with more to come. huggingface.co/blog/ggml-org…

Tensor parallel loading in 🤗 Diffusers got a massive upgrade 🔥 On Flux.2-Dev DiT, with a TP degree of 4 (A10G): • 30.4s → 12.5s loading • 64.1 → 6.8 GB peak CPU memory/rank ~2.4× faster loading. ~89% less CPU memory. Check it out: huggingface.co/docs/diffuser…

OpenAI launched dots. Here's how to build your own dots-style assistants with: > Pi as the harness > Custom gateway to connect to Telegram > Your choice of models (local, open, etc). > Google Workspace skills Run them on a VPS, Desktop, DGX, etc. Have fun!!

it's decision model season 🍂 today, we open sourced clef and clef-flash, our first homegrown models — both now available on @cloudflare workers ai! demo / leaderboard: clef-evals.workers-ai-mle.wo… blog: blog.cloudflare.com/clef-dec…

Introducing JEV-27B-VL. We believe it is the world’s first open-weight, near-SOTA multimodal decision model — built with AutoTrust AI’s Blocks of Experts (BoE) recipe. It does not just describe what it sees. It turns visual state into calibrated action probabilities, then selects what to do next. In this demo, it sees Mario’s position, movement, obstacles, and timing — then chooses whether to run, jump, or do both. See → decide → act.

🚀 Excited to announce that we are teaming up with @os4science to fund the software stack behind open source science! 🚀 Open science models do not run without open source software, but they rarely get the funding or credit they deserve. We're trying to change that 🤗🤗🤗

huggingface_hub v2.1.0 is out! ↻ Jobs: retries, rerun, reschedule & exposed ports 📊 Track your ZeroGPU quota ⚡ Up to 17× faster HfFileSystem downloads with hf_xet 🧠 Deploy Inference Endpoints from tested recipes pip install -U huggingface_hub github.com/huggingface/huggi…

Alert: Cloudflare just dropped a Jev alternative on Hugging Face 🚨 (Apache 2.0) huggingface.co/Cloudflare/cl…

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 hosted on workers ai or grab the weights from @huggingface, because we open-sourced it too. blog.cloudflare.com/clef-dec…

To celebrate scaling Microduck production, we scaled... the duck. Meet MEGADUCK at Open Together, Friday Oct 16 in SF - details below. How big do you think it is?

1/ Excited to release The ultimate guide to multi-harness RL The same model behaves differently in every agent harness. So we built an open way to train any model with RL on any task set, inside the harnesses people actually use, like Claude Code, Codex, and OpenCode, without changing a single line of harness or training code. Trained across four harnesses, LFM2.5-2.6B went from 42% to 54% with 31% fewer tool calls. 🧵

MCPs are making quite the comeback! You can now use them to bring your data to hf.co/chat?mode=ml-intern

You can now train an open model with RL inside claude code, codex, opencode, pi, or any harness you need. All at once, with the harness runs unmodified. @adithya_s_k added a capture proxy to openenv. it sits between the harness and the model and records the exact token ids and logprobs of every call. to the harness it's just another model provider. harbor supplies the tasks and sandboxes, trl trains with async GRPO. it's worth doing because the harness completely changes what the model learns. the same LFM2.5-2.6B weights solve 62% of held-out tasks in mini-swe-agent and 33% in claude code. after RL across four harnesses the average goes from 42% to 54%, and claude code from 33% to 49%. there are three envs you can try in the browser and a training script for each. guide: huggingface.co/spaces/Adithy…

Robots are going to eat the next 10 years 💯 We're making robotics on the Hugging Face Hub easier to use, eg: soon you'll be able to send any community-made policy to your Microduck. Can you just imagine the fun we're going to have?

Every company should be fine-tuning its own models on its own, proprietary business data. We've made this a LOT easier by adding MCP data sources in ML Intern, running on top of private, open models! Try it now: hf.co/chat?mode=ml-intern

I'm giving away $1,000 of HF credits 🤗 $10 x 100 people, I want you to experience this: > having your agent launch hundreds of HF Jobs > run ~33M tokens of DeepSeek V4.1 Flash > deploy your own dedicated Qwen3.8 27B endpoint > Agents + Blender on a GPU because it's 🔥 Just reply with your HF username 👇