Hunyuan-T1-20250711: benchmark results
Tencent's Hunyuan-T1 (July 11, 2025 checkpoint), a Hybrid-Transformer-Mamba Mixture-of-Experts reasoning model activating 52 billion parameters via dynamic expert routing, built on the TurboS foundation. Provider: Tencent. Released 2025-07-11. Access: API.
Aggregate Rating 1670 ± 37, rank #277 of 1635 rated models, from 85 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenCompass Language - Instruction Following - Chinese (CompassBench 2507) | 83.1 | Score (%) | 96.3 |
| OpenCompass Agent - Hard Multi-Turn (CompassBench 2507) | 50 | Score (%) | 91.7 |
| OpenCompass Reasoning - Common - English (CompassBench 2507) | 75.6 | Score (%) | 91.7 |
| SuperCLUE General (July 2025) - Hallucination Control | 83.39 | Score | 91.3 |
| OpenCompass Code - Competition - Chinese (CompassBench 2507) | 59.4 | Score (%) | 88.9 |
| OpenCompass Knowledge - Humanities - English (CompassBench 2507) | 87 | Score (%) | 88.9 |
| OpenCompass Code - Competition - English (CompassBench 2507) | 60.9 | Score (%) | 87 |
| SuperCLUE General (July 2025) - Precise Instruction Following | 42.73 | Score | 87 |
| OpenCompass LLM - Code (CompassBench 2507) | 43.8 | Score (%) | 85.2 |
| OpenCompass LLM - Code - Chinese (CompassBench 2507) | 43.2 | Score (%) | 85.2 |
| OpenCompass LLM - Code - English (CompassBench 2507) | 44.3 | Score (%) | 83.3 |
| SuperCLUE General (July 2025) - Math Reasoning | 68.7 | Score | 81.5 |
Interactive version: theaggregate.ai/model?slug=hunyuan-t1-20250711 · How It Works · Data refreshed daily, snapshot 2026-11-05.