A 200-line ReAct harness lets open models race to fix 3 real bugs
MaziyarPanahi · x · 2026-07-25
A real bug-fixing setup shows how multiple open models can be raced against the same repository issues.
- The author used 3 hidden bugs in a real repo, with pytest as the judge.
- Local models ran on a Mac Studio: Bonsai-27B, Inkling, and gemma-4-26B, served via llama.cpp + MLX.
- A separate stronger model, Kimi K3, was called through OpenRouter until the weights become public.
- The agent loop is a roughly 200-line ReAct harness that can read / edit / runtests and talk to any OpenAI-compatible endpoint.
The post frames this as a model-agnostic, practical way to compare coding agents on the same debugging task.
Related event: Open-Source Models Fix Real-World Bugs in ReAct Framework Test(2 posts)→
More from coding & agent
- Dual RTX 3090 owner asks which local LLM stack actually works today — ruffus_or · 2026-07-25
- ECC is a harness layer for Claude Code, Codex and Cursor agents — affaan-m · 2026-07-25
- Andrew Ng’s aisuite offers one interface for multiple GenAI providers — andrewyng · 2026-07-25
- AFK Pilot links Grok Build in VS Code to your phone with no tunnel or config — PawelHuryn · 2026-07-25
- Kimi K3 Coding Review: Is the Subscription Better Than the Pricey API? — agentcubed · 2026-07-25
- Four open models fix the same three real bugs, with a 27B model 14× faster — MaziyarPanahi · 2026-07-25