Log inSign up
Log inSign up
Roger Creus Castanyer
402 posts
Roger Creus Castanyer profile banner
@creus_roger

Roger Creus Castanyer

@creus_roger
Maximizing the unexpected return. SR @Google. PhD @Mila_Quebec / @VmaxAI | Prev: @MSFTResearch @Ubisoft @la_UPC
Montréal, Québec
roger-creus.github.io
Joined April 2021
799 Following
1,023 Followers
RepliesRepliesRepostsRepostsMediaMediaArticlesArticles

Roger Creus Castanyer’s posts

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·US TIDA·Ads Info·© 2026 X Corp.
  • Pinned
    @creus_roger
    Roger Creus Castanyer
    @creus_roger
    Sep 29
    Article cover image
    Article
    I let GPT-6 Astra run its own research. It beat Craftax and reached the bottom of NetHack.
    TL;DR: For three weeks, GPT-6 Astra did the research. It read its own failed games, rewrote the prompts, tools, and memory of the Astra agents that were playing, and launched new games to test each
    17
  • @creus_roger
    Roger Creus Castanyer
    @creus_roger
    Nov 10
    @VmaxAI 🤝 @modal Very happy to finally share this! We worked with @modal to make coding agents better at using the Modal Python SDK with our in-house self-play algorithms, and the results were 🔥🔥 All the details in the blog post below! ⬇️
    @VmaxAI
    Vmax
    @VmaxAI
    Nov 10
    1/ As a teaser for Campaign, Vmax’s infrastructure for automated task generation, we are sharing work we did earlier this year for @modal using self-play to generate tasks that target weaknesses of models when using the Modal-SDK.
    00:00
    5
  • @creus_roger
    Roger Creus Castanyer
    @creus_roger
    Oct 28
    "including training in games [...] strategies learned there transfer to research mathematics and other fields" we also showed with GlyphBench that training LLMs in games can yield general reasoning gains! x.com/VmaxAI/status/…
    @imjustnewatai
    imjustnewatai
    @imjustnewatai
    Oct 28
    Rumors I’ve been hearing. I’ll start with OpenAI, then Anthropic. OpenAI: The internal model’s reasoning jump comes from more aggressive RL, including training in games and simulated environments. Strategies learned there transfer to research mathematics and other fields.
  • @creus_roger
    Roger Creus Castanyer
    @creus_roger
    Oct 23
    🚨 Excited to release GlyphBench, our new paper on games, RL, and LLM reasoning! big, big thanks to @MavorParker, @matthewjsargent, @Cote_Marc, @GlenBerseth & @pcastr and shoutout to @wwwjim for the beautiful blog design! preprint, code, and blog post are all public! ✨
    @VmaxAI
    Vmax
    @VmaxAI
    Oct 23
    1/ 🎮 Can training on games improve LLM reasoning? Introducing GlyphBench: 360+ game tasks for language-model RL. We use it as a benchmark, a research tool, and a post-training suite. 🧵⬇️
    00:00
    3
  • @creus_roger
    Roger Creus Castanyer
    @creus_roger
    Oct 1
    Blazing fast agentic RL with KV-streams 🚀 It's been great to be part of this big collaboration 🤝🤝
    @emilianopp_
    Emiliano Penaloza
    @emilianopp_
    Sep 30
    Want to train your SWE agents 2× faster 🏎️💨? We introduce KV-streams, which, rather than flushing the KV cache on agentic compaction, we preserve it and keep generating. This avoids the re-prefill cost on both the inference engine and the trainer. Unlike other works that
    00:00