Ouais, cela semble avoir été juste compte tenu de l'agent unique contre grok 4 agent multi lourd.
Jimmy Apples 🍎/acc
Jimmy Apples 🍎/acc11 juil. 2025
Hearing a few whispers now from birds that internal evals are having gpt5 a tad over grok 4 Heavy. Evals only tell one side to a model however, curious to see if we get any major agentic or other improvements.
25,13K