The authors ran 1,886 npm releases of Claude Code, Codex, Gemini CLI, Qwen Code and OpenCode in sandboxes, sent each the same prompt, and recorded the first request each agent made to its main model. They call everything the agent adds…
In the News: October 29, 2026 (Extra)
A study of 1,886 coding-agent releases finds tool definitions make up 77.4% of Claude Code's first request and that fewer than half of prompt changes explain themselves.
Extra edition
Machine-readable
Download Markdown