Zhoujun (Jorge) Cheng

🤯What we know about RL for reasoning might not hold outside math and code? We revisit established findings on RL for LLM reasoning on six domains (Math, Code, Science, Logic, Simulation, Tabular) and found that previous conclusions drawn on math and code are surprisingly…

Proud to introduce Group Sequence Policy Optimization (GSPO), our stable, efficient, and performant RL algorithm that powers the large-scale RL training of the latest Qwen3 models (Instruct, Coder, Thinking) 🚀 📄 huggingface.co/papers/2507.18…

Wrapped up a SWE-Perf website redesign using Qwen3-Coder on AnyCoder (huggingface.co/spaces/akhaliq…). The process was incredibly fast and great! One question for Qwen devs, though: did you pretrain a secret love for the color purple into the coder's persona? 😉

Countless times of iterations for cooking it, but the process is satisfying. I still believe we can pour more data in each stage if we have more hands so the potential is unlimited and scaling law hasn’t hit the wall yet! Towards Digital Agents🤖 We are already on the way.

Excited to bring Qwen3-Coder into the browser and terminal world! Building the scaffolding and environments for this big guy to play and learn is tough but incredibly "rewarding". Agentic coding and browsing are arguably the two most important skills for digital agents: they…

Apart from the performance, it’s pure entertainment just watching Qwen3‑Coder build Qwen Code all by itself. Agentic coding is really something: it explores, understands, plans, and acts seamlessly. Honored to be “in the game”—even if my entire work so far is smashing the Enter…

🥰

We’re open-sourcing the pre-training code for Phi4-mini-Flash, our SoTA hybrid model that delivers 10× faster reasoning than Transformers — along with μP++, a suite of simple yet powerful scaling laws for stable large-scale training. 🔗 github.com/microsoft/Arch… (1/4)

🔥 LLMs can fix bugs, but can they make your code faster? We put them to the test on real-world repositories, and the results are in! 🚀 New paper: "SWE-Perf: Can Language Models Optimize Code Performance on Real-World Repositories?" Key findings: 1️⃣ We introduce SWE-Perf, the…

🚨 70 million US workers are about to face their biggest workplace transmission due to AI agents. But nobody asks them what they want. While AI races to automate everything, we took a different approach: auditing what workers want vs. what AI can do across the US workforce.🧵

If you are at #icml25 and are interested in RL algorithms, scaling laws for RL, and test-time scaling (& related stuff), come talk to us at various poster sessions (details ⬇️). We are also presenting some things at workshops later in the week, more on that later.

🚀 Thrilled to announce Dream-Coder 7B — the most powerful open diffusion code  LLM to date.

👇this nice guy❤️will help us present CodeI/O (arxiv.org/abs/2502.07316) at Oral session 6A Applications in Agents and Coding, Thu 17 Jul 4 p.m. — 4:15 p.m. PDT. Take a look if you are there and feel interested.

I still find it mysterious whether and how intelligence and capabilities transfer between domains and skills - from meta learning during early days to more recent question like whether solving maths helps writing a good essay. Sometime I feel a bit pessimistic given not enough…

Prompting is our most successful tool for exploring LLMs, but the term evokes eye-rolls and grimaces from scientists. Why? Because prompting as scientific inquiry has become conflated with prompt engineering. This is holding us back. 🧵and new paper: arxiv.org/abs/2507.00163

🚀 Hello, Kimi K2! Open-Source Agentic Model! 🔹 1T total / 32B active MoE model 🔹 SOTA on SWE Bench Verified, Tau2 & AceBench among open models 🔹Strong in coding and agentic tasks 🐤 Multimodal & thought-mode not supported for now With Kimi K2, advanced agentic intelligence…

If you're attending #ICML2025, check out our 💭 Agent Workflow Memory for online adaptive agents: Jul 17 4:30-7pm @ West Hall 🔎 RAGGED for designing scalable and stable RAG systems: Jul 16 11:00-13:30 @ East Hall Computer Use Agent Workshop on Jul 19 🌐 "Universal Retrieval for…

Our Coconut work is now accepted in COLM'25. Thanks all the reviewers for the support and constructive feedbacks!

🚀 Check out our recent work Afterburner: Reinforcement Learning demonstrating super powerful self-improving code efficiency optimization! 💻✨

🤔 Ever wonder where reinforcement learning actually boosts (or hurts) LLM’s reasoning capabilities? Meet SPARKLE—a new analysis framework that dissects gains from RL in planning, knowledge integration, and subproblem solving. 📄 Paper: arxiv.org/abs/2506.04723 🌐 Project:…

ASI is now accepted to @COLM_conf #COLM2025! 🍁 🔗 arxiv.org/abs/2504.06821