NLCO - Set-M - Accuracy: leaderboard
Metric: Accuracy: feasible and matching the solver-optimal objective (%). Source: arxiv.org. Saturation forecast: Around December 2026. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Flash (High) | 76.8 |
| 2 | GPT-5.1 (Medium) | 72.8 |
| 3 | Grok 4.1 Fast (Reasoning) | 70.9 |
| 4 | O4 Mini (High) | 70.7 |
| 5 | Claude Sonnet 4.5 (Thinking) | 70.4 |
| 6 | DeepSeek V3.2 (Thinking) | 69.1 |
| 7 | Qwen 3 235B A22B 2507 Instruct | 54.9 |
| 8 | DeepSeek V3.2 (Non-reasoning) | 47.7 |
| 9 | Qwen 3 14B (Reasoning) | 42.4 |
| 10 | QwQ-32B | 39.1 |
| 11 | MiMo-V2-Flash (Non-reasoning) | 37.4 |
| 12 | Ministral-3-14B-Instruct-2512 | 25.4 |
| 13 | Llama 4 Maverick Instruct | 19 |
| 14 | Qwen 3 14B (Non-reasoning) | 17.5 |
Interactive version: theaggregate.ai/benchmark?slug=nlco-set-m-accuracy · How It Works · Data refreshed daily, snapshot 2026-09-25.