ProgramBench

The AI receives a working program and its documentation, then builds a replacement from scratch without the original source code, internet access or decompilation. The score is the percentage of 200 programs that pass every behavioral test. We record each model's best published mini-SWE-agent result, including higher reasoning efforts where available. Partial test-pass rates and almost-solved programs do not count toward this score. Equal scores share a rank here; the official board also uses partial progress to break ties. Higher is better.

Claude Opus 5 by Anthropic holds the best ProgramBench result among tracked models, at 4.5% (released Jul 24 2026). Higher scores are better on this benchmark.

1. Claude Opus 5 (Anthropic) — 4.5%; 2. Muse Spark 1.3 (Meta) — 2.5%; 3. GPT-5.6 Sol (OpenAI) — 1%; 4. GPT-5.5 (OpenAI) — 0.5%; 4. Gemini 3.6 Flash (Google) — 0.5%.

20 tracked models have a published ProgramBench score. Scores are gathered from ProgramBench, which runs the benchmark and publishes the full field, and are kept in step with its board.