NLCO - Set-M - Feasibility Rate: leaderboard
Metric: Average feasibility rate: output satisfies every constraint (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.1 (Medium) | 98 |
| 2 | Gemini 3 Flash (High) | 97.5 |
| 3 | DeepSeek V3.2 (Thinking) | 96.8 |
| 4 | Claude Sonnet 4.5 (Thinking) | 96.7 |
| 5 | O4 Mini (High) | 96.5 |
| 6 | Grok 4.1 Fast (Reasoning) | 93.6 |
| 7 | Qwen 3 235B A22B 2507 Instruct | 87.6 |
| 8 | DeepSeek V3.2 (Non-reasoning) | 86.1 |
| 9 | Qwen 3 14B (Reasoning) | 79.1 |
| 10 | QwQ-32B | 73.8 |
| 11 | Llama 4 Maverick Instruct | 65.1 |
| 12 | MiMo-V2-Flash (Non-reasoning) | 62.8 |
| 13 | Ministral-3-14B-Instruct-2512 | 58.7 |
| 14 | Qwen 3 14B (Non-reasoning) | 58.1 |
Interactive version: theaggregate.ai/benchmark?slug=nlco-set-m-feasibility-rate · How It Works · Data refreshed daily, snapshot 2026-09-25.