• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
  • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    Leanstral 1.5 beat GPT-6 Astra on ArXivLean, a user reports

    The user describes Leanstral as a 6-billion-parameter model and suspects the displayed token count is incorrect.

    T(
    LA
    3 Sources, 19d ago, first seen 19d ago

    TLDR

    A user says Leanstral 1.5 beat GPT-6 Astra on ArXivLean with “a lot of grinding” and “a tiny bit of K3 orchestration.” They describe it as a 6-billion-parameter model, while cautioning that they think the displayed token count is erroneous.

    Combined views

    34.5K

    3 Sources, first seen 19d ago

    389 likes

    Combined views

    34.5K

    3 Sources, first seen 19d ago

    389 likes
    16 comments
    99 saves
    11 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    16 comments
    99 saves
    11 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    3 Sources

    @AlbertQJianghttps://matharena.ai/?comp=arxivlean--june&view=problem A lot of grinding from Leanstral 1.5 with a tiny bit of K3 orchestration get to beat GPT-6 Astra on ArXivLean. I think #tokens is counted erroneously here but yes a 6B model does yap. Great internship work by Matéo Pirio Rossignol mentored by @roman_soletskyi!
    @scaling01Leanstral + K3 mogged GPT-6-Astra with >10M tokens now imagine what GPT-6-Astra with 130B tokens does
    @teortaxesTexThis sounds a lot but if it's 10M tokens primarily from Leanstral, that's not a great deal. Leanstral is 6B active. Acceptable price for going beyond Astra. A stronger composite of this shape could probably get 90%. The question is how much work the coordinator did

    3 Sources

    @AlbertQJianghttps://matharena.ai/?comp=arxivlean--june&view=problem A lot of grinding from Leanstral 1.5 with a tiny bit of K3 orchestration get to beat GPT-6 Astra on ArXivLean. I think #tokens is counted erroneously here but yes a 6B model does yap. Great internship work by Matéo Pirio Rossignol mentored by @roman_soletskyi!
    @scaling01Leanstral + K3 mogged GPT-6-Astra with >10M tokens now imagine what GPT-6-Astra with 130B tokens does
    @teortaxesTexThis sounds a lot but if it's 10M tokens primarily from Leanstral, that's not a great deal. Leanstral is 6B active. Acceptable price for going beyond Astra. A stronger composite of this shape could probably get 90%. The question is how much work the coordinator did