external benchmark
ExploitBench leaderboard
ExploitBench v8-bench. Every model the catalog carries a published ExploitBench value for, ranked by that value.
Categoryexternal
MeasureCapability coverage percentage over 16 flags
TasksV8 exploit synthesis runs
DifficultyBrowser exploitation and cybersecurity
Published byExploitBench
A cybersecurity benchmark for evaluating LLM agents on full-control V8 exploit synthesis using 16 measured exploit capability flags.
ExploitBench ranking
8 models with a published ExploitBench value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Capability coverage percentage over 16 flags |
|---|---|---|---|
| 1 |  GPT-6 Astra | OpenAI | 100 |
| 2 |  Claude Mythos 5 | Anthropic | 78 |
| 3 | GPT-5.6 Sol | OpenAI | 74 |
| 4 | Claude Mythos Preview | Anthropic | 69 |
| 5 |  GLM-5.3 | Z.AI | 54 |
| 6 | GPT-5.6 Terra | OpenAI | 53 |
| 7 | GPT-5.6 Luna | OpenAI | 33 |
| 8 |  Kimi K3 | Moonshot AI | 32 |
Evidence key: Observed
Rows are ordered by the value ExploitBench published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.