CCR-Bench - Industrial Hard Satisfaction: leaderboard
Metric: Hard satisfaction rate (0-1, times 100): share of instructions whose constraints are all satisfied, on CCR-Bench's 40 industrial (medical-scenario) instructions with expert-defined constraints from 15 constraint types, temperature 0, model-plus-rule scoring averaged over 10 assessments; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 41.5 | #145 |
| 2 | DeepSeek R1 0528 | 31.5 | #217 |
| 3 | Qwen 3 32B (Thinking) | 24.7 | #424 (Qwen 3 32B) |
| 4 | O3 Mini | 24.2 | #266 |
| 5 | Qwen 3 32B (Non-reasoning) | 15.8 | #424 (Qwen 3 32B) |
| 6 | QwQ-32B | 15.2 | #410 |
| 7 | GPT-4.1 | 13.5 | #240 |
| 8 | DeepSeek V3 (0324) | 10.2 | #332 |
| 9 | Qwen 2.5 72B Instruct | 8 | #436 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=ccr-bench-industrial-hard-satisfaction · How It Works · Data refreshed daily, snapshot 2026-10-11.