Same model, 17 points apart: OpenCode hits 82% on 21 MCP servers vs Claude Code's 65%
alejxb · reddit · 2026-11-01
Running the same model (Claude Haiku 5.5) against the same 21 MCP servers through two different coding agents, only the agent differed: OpenCode answered 82.0% of tasks correctly while Claude Code managed 65.4% — a 17-point gap. The chart also shows Codex (Luna) failing more often. Takeaway: the agent harness can matter more than the model for MCP tool-use success.
More from coding & agent
- UIUC, Meta, Amazon & DeepMind release 135-page survey charting agentic reasoning for LLMs — mdancho84 · 2026-11-01
- 10 open-source GitHub repos to build an AI-powered second brain — Shruti_0810 · 2026-11-01
- When AI writes almost all code: Pragmatic Engineer on the birth of Product Engineering — carlcarrie · 2026-11-01
- ai-engineering-hub tutorials repo surges past 39.7k GitHub stars — patchy631 · 2026-11-01
- AiSOC: open-source AI security operations center with LLM-agent triage hits 5.8k stars — beenuar · 2026-11-01
- HKUDS' open-source AI tutor DeepTutor 1.7 ships topic-to-video lessons and in-note Q&A — YiMaTweets · 2026-11-01