I'm not sure how this makes sense when Claude models aren't even coding specific: Haiku, Sonnet, Opus are the exact same models you'd use for chat or (with the recent Mythos) bleeding edge research.
replyAnthropic models and training data is optimized for coding use cases, this is the difference.
replyOpenAI on the other hand has different models optimized for coding, GPT-x-codex, Anthropic doesnt have this distinction
But they detect it under the hood and apply a similar "variant", as API results are not the same than on Claude Code (that was documented before by someone).
reply