Pick a winner in each dimension — change your vote anytime.
Pricing and capabilities synced from the OpenRouter catalogue. Scores are editorial (0–10).
Choose GPT-5.3-Codex if you want codex CLI users and dedicated coding agents. Choose Kimi K2.7 Code if you want cheap, capable open-weight coding agents.
Editorially it's close: each model leads in 2 of our five dimensions. On price, GPT-5.3-Codex runs about $5.42 per 1M tokens (blended) and is proprietary; Kimi K2.7 Code is about $1.35 and open-weight.
The model picks the moves; the agent runs the loop, the tools, and the guardrails. Once you've chosen a model, see which agent gets the most out of it.