- MMLU PRO 87.2
- GPQA Diamond 69.3
- C-SimpleQA 67.9
- DROP F1 93.1
- Zebra Logic 79.6
- MATH-500 96.2
- AIME 2024 78.2
- LiveCodeBench 64.9
- CEval 91.8
- CMMLU 90
- ArenaHard 91.9
- CFBench 81
- CELLO 76.4
- T-Eval 68.8
Hunyuan-T1
Hunyuan-T1 places in the top three on DROP.
16
Benchmarks with a score
0
Ranked first, out of five or more models
1
In the top three
Tencent Mar 2025 Hunyuan-T1
Closed weights Flagship
135.7
#222 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
- ClawProBench 34.7% #49 /49
- MMLU-Pro 87.2% #8 /167
- C-Eval 91.8% #10 /46
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MMLU-Pro | 87.2% | #8 of 167 | Lab model card | |
| DROP | 93.1% | #2 of 41 | Lab model card | |
| Arena-Hard-Auto | 91.9% | #7 of 39 | Lab model card | |
| C-Eval | 91.8% | #10 of 46 | Lab model card | |
| CMMLU | 90% | #6 of 17 | Lab model card | |
| LMArena (Elo) | 1,387 | #148 of 335 | Official leaderboard | |
| CFBench | 81% | #1 of 2 | Lab model card | |
| Chinese SimpleQA | 67.9% | #9 of 13 | Lab model card | |
| CELLO | 76.4% | #1 of 1 | Lab model card |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MATH-500 | 96.2% | #33 of 70 | Lab model card | |
| AIME 2024 | 78.2% | #47 of 88 | Lab model card | |
| GPQA Diamond | 69.3% | #186 of 302 | Lab model card | |
| ZebraLogic | 79.6% | #18 of 23 | Lab model card |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LiveCodeBench | 64.9% | #17 of 54 | Lab model card |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ClawProBench | 34.7% | #49 of 49 | Official leaderboard | |
| T-Eval | 68.8% | #2 of 2 | Lab model card |
Its lab's numbers put it at 135.7, number 222 of 404
The Self-Reported Index fits one score to the 11 capability results Tencent reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 133.4 to 137.9. How we compute it.
- MATH-500 96.2
- DROP 93.1
- Arena-Hard-Auto 91.9
- C-Eval 91.8
- CMMLU 90
- MMLU-Pro 87.2
- ZebraLogic 79.6
- AIME 2024 78.2
- GPQA Diamond 69.3
- Chinese SimpleQA 67.9
- LiveCodeBench 64.9
Its lab's number next to a leaderboard
Where the benchmark's own leaderboard, or Epoch AI's run, also has this model. Gaps mostly come from different harnesses, effort levels and task subsets. Every pair.
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.