# Anthropic OSS-Fuzz Any-Crash Rate (Anthropic OSS-Fuzz crash)

> Share of evaluated OSS-Fuzz entry points where the model produced at least a crash.

Canonical page: https://benchlm.ai/benchmarks/anthropicossfuzzanycrash

- Category: [Agentic](/agentic)
- Last updated: September 27, 2026

## About Anthropic OSS-Fuzz crash

- Year: 2026
- Tasks: Approximately 830 OSS-Fuzz entry points
- Format: Any-crash rate
- Difficulty: Vulnerability discovery
- Paper: [Claude Fable 5 and Claude Mythos 5](https://www.anthropic.com/news/claude-fable-5-mythos-5)

This is Anthropic's internal OSS-Fuzz-derived protocol over roughly 830 entry points, not a public OSS-Fuzz leaderboard. The row is display only and preserves the provider-specific protocol name.

Anthropic OSS-Fuzz crash is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (1 models)

| Rank | Model | Creator | Score |
|------|-------|---------|-------|
| 1 | [Claude Mythos 5](/models/claude-mythos-5) | Anthropic | 80.0% |

## FAQ

### What does Anthropic OSS-Fuzz crash measure?

Share of evaluated OSS-Fuzz entry points where the model produced at least a crash.

### Which model scores highest on Anthropic OSS-Fuzz crash?

Claude Mythos 5 by Anthropic currently leads with a score of 80.0% on Anthropic OSS-Fuzz crash.

### How many models are evaluated on Anthropic OSS-Fuzz crash?

1 AI models have been evaluated on Anthropic OSS-Fuzz crash on BenchLM.

### Does Anthropic OSS-Fuzz crash affect BenchLM's overall score?

Not directly. Anthropic OSS-Fuzz crash is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.