❄️Introducing Absolute Zero Reasoner: Our reasoner learns to both propose tasks that maximize learnability and improve reasoning by solving them, entirely through self-play—with no external data! It overall outperforms other "zero" models in math & coding domains. 🧵 1/

RT @BanghuaZ: Beyond prompt / context engineers, we’re seeing the rise of environment engineers, experts who build high-quality RL environm….

RT @zzlccc: With just a few lines of code, Feng’s (@fengyao1909) suggested fix—applying importance sampling on the behavior policy—resolved….

RT @Gradient_HQ: Reinforcement Learning is the future tense of intelligence. Echo is how it scales. Echo is Gradient’s distributed RL fram….

RT @willccbb: the easiest way to get hired at @primeintellect for research is to just make it very clear that you're already doing excellen….

RT @kuchaev: We are excited to release Nvidia-Nemotron-Nano-V2 model! This is a 9B hybrid SSM model with open base model and training data.….

RT @gm8xx8: NVIDIA Nemotron-Nano v2. Models: 12B Base, 9B Reasoning, 9B Base.- Arch: Hybrid Mamba2–Transformer (128K ctx, 4 attn layers).-….

RT @chatgpt21: It looks like Andrew Garfield will play Sam Altman in the Open AI movie coming to Amazon MGM. I think @apples_jimmy deserves….

RT @gensynai: Coming soon

LLMs as internet/knowledge base, no need for external tools. Reminiscent of older work from AI2/UW, Rainer and CRYSTAL

RT @hwchung27: After a great time at OpenAI, we (@EdwardSun0909, @_jasonwei) recently joined @Meta Superintelligence Labs. The first month….

RT @DimitrisPapail: Thinking Less at test-time requires Sampling More at training-time!. GFPO is a new, cool, and simple Policy Opt algorit….

RT @basvanopheusden: A few weeks ago, I started a new job at @OpenAI. I wrote a document about my interview process and recommendations for….

RT @arcprize: "I've updated my AGI timeline.". One year later, @dwarkesh_sp and @fchollet meet on camera again. Both of them have shifted….

Nice empirical paper investigating all your bag of tricks in reasoning LLMs.

RT @GXiming: 🚀 How far can RL scaling take LLMs?.Drop ProRLv2! 🔥We keep expanding LLM’s reasoning boundaries through 3,000+ RL steps over 5….

RT @canondetortugas: Announcing the first workshop on Foundations of Language Model Reasoning (FoRLM) at NeurIPS 2025!. 📝Soliciting abstrac….

RT @ericzelikman:

RT @interconnectsai: GPT 5 Launch Party w/ Will Brown & Swyx

RT @fengyao1909: Failing on 𝐥𝐚𝐫𝐠𝐞-𝐬𝐜𝐚𝐥𝐞 𝐑𝐋 with VeRL?. ⚠️ Mixing inference backend (𝐯𝐋𝐋𝐌/𝐒𝐆𝐋𝐚𝐧𝐠) with training backends (𝐅𝐒𝐃𝐏/𝐌𝐞𝐠𝐚𝐭𝐫𝐨𝐧) 𝐬𝐞𝐜….

RT @zzlccc: Excited to see GDM proposing Game Arena to measure the model capabilities. Let’s also scale the environments for agent RL with….