Benchmark
ExploitBench
ExploitBench is a cybersecurity benchmark that evaluates a model's ability to discover and exploit software vulnerabilities, reported as the fraction of…
- Modality
- text
- Categories
- safety, agents, code
- Openness
- unknown
- Source
- llm_stats
- Reported scores
- 5
Reported scores
llm_stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| Claude Fable 5 | Anthropic | 0.78 | 2026-06-09 | evidence |
| GLM-5.3 | Z.ai | 0.544 | 2026-08-14 | evidence |
| GPT-5.6 Luna | OpenAI | 0.332 | 2026-07-09 | evidence |
| GPT-5.6 Sol | OpenAI | 0.735 | 2026-07-09 | evidence |
| GPT-5.6 Terra | OpenAI | 0.529 | 2026-07-09 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.