IFMTBench - Multi-Constraint: leaderboard
Metric: IF_Score (%): hard constraints checked by deterministic verifiers gate the item, soft constraints are scored by a gpt-oss-120b rubric judge, mean over items, on the 2,838 multi-constraint items, translation requests into seven languages with instructions paraphrased in all seven, officially recommended decoding with non-thinking mode for open models; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 84.53 |
| 2 | Qwen 3.5 27B (Non-reasoning) | 78.81 |
| 3 | Hy-MT2-30B-A3B | 75.8 |
| 4 | Gemma 4 26B A4B | 73.89 |
| 5 | Gemma 4 31B | 72.64 |
| 6 | Qwen 3.6 35B A3B (Non-reasoning) | 71.43 |
| 7 | Qwen 3.5 35B A3B (Non-reasoning) | 70.32 |
| 8 | Gemma 4 E4B | 69.25 |
| 9 | Qwen 3.5 9B (Non-reasoning) | 64.28 |
| 10 | Hy-MT2-1.8B | 57.61 |
| 11 | Qwen 3.5 4B (Non-reasoning) | 57.23 |
| 12 | Gemma 4 E2B | 50.72 |
| 13 | Qwen 3.5 2B (Non-reasoning) | 24.6 |
| 14 | Qwen 3.5 0.8B (Non-reasoning) | 7.46 |
Interactive version: theaggregate.ai/benchmark?slug=ifmtbench-multi-constraint · How It Works · Data refreshed daily, snapshot 2026-10-07.