DecBench: leaderboard
Decompilation-exactness benchmark (Noelo Lab, UGA): share of binary functions recovered perfectly across structure, types, and recompilation. LLM agent backends are scored on the curated sample-set tier alongside traditional decompilers.
Metric: Exact Recovery (%). Source: decbench.com. Saturation forecast: Around January 2028. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra | 59.6 | #6 |
| 2 | Claude Opus 4.8 | 56.8 | #35 |
No result here: #4 Claude Opus 5.5, #7 GPT-6.1 Sol, #8 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=decbench · How It Works · Data refreshed daily, snapshot 2026-10-23.