Hy-MultiTurn - Constraint Synthesis: leaderboard
Metric: Importance-weighted score (%) on the 35 constraint synthesis dialogues, 209 long Chinese multi-turn dialogues (12 to 76 turns), three rolls per task averaged, three LLM judges (GPT-5.2, Qwen 3.6, DeepSeek-V4 Pro) averaged, a deterministic checker for precise execution; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 22 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Grok 4.5 (High) | 90 |
| 2 | GPT-5.5 (High) | 84.9 |
| 3 | GPT-5.4 (High) | 83.7 |
| 4 | Claude Opus 4.6 (High) | 83.4 |
| 5 | Kimi K3 (Max) | 80.2 |
| 6 | Hy3 (High) | 79.5 |
| 7 | Claude Opus 4.7 (High) | 74.9 |
| 8 | Kimi K2.6 | 68.8 |
| 9 | Doubao-Seed-2.1-Pro (High) | 65.3 |
| 10 | Hy3-preview (High) | 62.6 |
| 11 | Kimi K2.6 (Non-reasoning) | 61.5 |
| 12 | Seed 2.0 Pro (Thinking) | 59.7 |
| 13 | Kimi K2.5 (Non-reasoning) | 59.3 |
| 14 | Qwen 3.6 Plus (Thinking) | 56.9 |
| 15 | Hy3-preview (Non-reasoning) | 55.6 |
Interactive version: theaggregate.ai/benchmark?slug=hy-multiturn-constraint-synthesis · How It Works · Data refreshed daily, snapshot 2026-09-29.