LatentMD - Latent Failure Rate: leaderboard
Metric: Content-correct responses with a latent boundary failure (%). Source: arxiv.org. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemma 3 27B | 73.8 |
| 2 | Gemini 2.5 Flash | 49.9 |
| 3 | Qwen 3 32B | 46.1 |
| 4 | GPT-4o | 43.3 |
| 5 | Llama 3.1 70B | 35.7 |
| 6 | Claude Sonnet 4 | 30.6 |
| 7 | Llama 3.1 8B | 14.5 |
| 8 | Gemma 2 9B | 1.4 |
Interactive version: theaggregate.ai/benchmark?slug=latentmd-latent-failure-rate · How It Works · Data refreshed daily, snapshot 2026-09-20.