A 30-game battle royale across eleven LLMs, $482 of inference, and one finding that should change how you read model benchmarks.
  • SmokeInFog@midwest.socialEnglish
    2·
    21 days ago

    I feel like this article is a bunch of non-sequiturs strung together. He even claims that his LLM agent battle is something “a routing customer actually cares about.”

  • Em Adespoton@lemmy.ca
    2·
    21 days ago

    Either is fine. I’d just send it a wall of text and it would hit its token limit and stall.

    Unless it was running towards me to protect me; in that case, I’d want Claude.