4 ms·Presumably because GLM 4.5 or Qwen3 comparisons would clobber them in eval scores.by bigyabai 1y agoPresumably because GLM 4.5 or Qwen3 comparisons would clobber them in eval scores.quotemstr 1y agoAnd don't require KYC crap to predict next tokenconradkay 1y agoYou can check the same evals OpenAI used for those models Hint: unclobbered