LogiMed-RoB - Domain Logical Fidelity: leaderboard
LogiMed-RoB: whether each domain-level verdict follows logically from the atomic judgements beneath it, averaged over the five RoB 2.0 domains. Higher is better. Source: arxiv.org/abs/2609.11185.
Metric: Fidelity (%). Source: arxiv.org. Saturation forecast: Around February 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 87.61 | #64 |
| 2 | Gemini 3 Flash | 83.9 | #103 |
| 3 | DeepSeek V3.2 | 79.99 | #210 |
| 4 | Claude Sonnet 4.5 | 79.22 | #148 |
| 5 | GPT-5.1 | 78.65 | #145 |
| 6 | Llama 4 Maverick | 61.54 | #486 |
| 7 | Llama 3.3 70B | 60.1 | #578 |
| 8 | Qwen 3 32B | 51.39 | #449 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=logimed-rob-domain-logical-fidelity · How It Works · Data refreshed daily, snapshot 2026-11-05.