ExploitBench leaderboard
- GPT-5.5 leads
- 47.4%
- models measured
- 8
- lead over #2
- +20.9 pts
- median
- 21.0%
Leader on ExploitBench · read October 21
On October 21, 2026, 8 of our catalog's models had a ExploitBench result, and OpenAI GPT-5.5 led with 47.4%.
Epoch AI collects its results; its benchmarking hub describes the tasks and links the authors.
Role in the index
We collect this benchmark and show it as supporting evidence. It is not part of the Vibe Coding Index, because it measures something other than coding and the index is about coding. Read the methodology
| # | Model | Result | Evidence | Coding index | Output $/Mtok |
|---|---|---|---|---|---|
| 1 | GPT-5.5OpenAI | 47.4% | 67.8 | $30.00 | |
| 2 | Claude Opus 4.7Anthropic | 26.5% | 51.1 | $25.00 | |
| 3 | Gemini 3.1 Pro PreviewGoogle | 26.1% | 35.3 | $12.00 | |
| 4 | Claude Sonnet 4.6Anthropic | 23.6% | 57.8 | $15.00 | |
| 5 | Kimi K2.6Moonshot AI | 18.4% | 44.2 | $3.41 | |
| 6 | GLM 5.1Z.ai | 18.1% | 44.0 | $4.40 | |
| 7 | Claude Haiku 4.5Anthropic | 13.7% | 49.8 | $5.00 | |
| 8 | MiniMax M2.7MiniMax | 13.3% | 28.4 | $1.20 |
Only models in our catalog are listed, and each appears once with its best published run. 197 rows were read from the source on October 21, 2026, and rows that match no catalog model are left out.