Log inSign up
Vals AI
1,858 posts
Vals AI profile banner
@ValsAI

Vals AI

@ValsAI
Public LLM Evaluation // vals.ai @8vc @BloombergBeta @pearvc @a16z @HRTVentures @nextladder
San Francisco, CA
vals.ai
Joined March 2024
272
Following
18K
Followers
RepliesRepliesRepostsRepostsMediaMedia

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·Ads Info·© 2026 X Corp.
  • Pinned
    @ValsAI
    Vals AI
    @ValsAI
    Sep 1
    Using Fable 5.1, we’ve elicited a decoding of a cipher that has been unsolved for 373 year, and listed as #28 on Klaus Schmeh’s (the most-published cryptology author in the world) “The Top 50 unsolved encrypted messages.” The solution? Almost embarrassingly simple in hindsight.
    Image
    Made with AI
    44
  • @ValsAI
    Vals AI
    @ValsAI
    10h
    SUPERHOT is another game in our public set of games. Astra plays surprisingly well. At around 2 hours and 7 minutes, it punches an NPC, steals their gun in mid air, then shoots them.
    Image
    00:00
    @ValsAI
    Vals AI
    @ValsAI
    10h
    Image
    03:01:05
    Astra began to saturate a computer-use eval we haven’t even released yet. Astra was not only able to kill several hostile mobs in Minecraft, but it made it all the way to building a Nether portal in less than three hours. We believe that if we ran it longer, it would beat the
    1
  • @ValsAI
    Vals AI
    @ValsAI
    10h
    Astra began to saturate a computer-use eval we haven’t even released yet. Astra was not only able to kill several hostile mobs in Minecraft, but it made it all the way to building a Nether portal in less than three hours. We believe that if we ran it longer, it would beat the
    Image
    00:00
    11
  • @ValsAI
    Vals AI
    @ValsAI
    10h
    GPT-6 Astra lands at #3 on the Vals RSI Index, ahead of GPT-5.6 Sol but behind Fable 5.1 and Opus 5.
    Image
    GIF
    2
  • @ValsAI
    Vals AI
    @ValsAI
    12h
    Today we are releasing Tax Agent Bench, an agentic benchmark evaluating LLMs’ ability to do professional-level US corporate tax research. It consists of 391 expert-written questions with each task evaluated with a rubric developed by tax practitioners.
    Image
    9
Advertisement
Advertisement