kimi-k2.5 is a thinking model that times out on complex prompts (>3min). qwen3.5-397b-a17b responds in ~2s with clean JSON output. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
kimi-k2.5 is a thinking model that times out on complex prompts (>3min). qwen3.5-397b-a17b responds in ~2s with clean JSON output. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>