Qwen2.5-Max
Qwen2.5-Max places in the top three on 2 benchmarks.
22
Benchmarks with a score
0
Ranked first, out of five or more models
2
In the top three
Qwen Jan 2025 Qwen2.5-Max
Closed weights Flagship
131.2
#155 of 256
Epoch Capabilities Index
- GPQA Diamond 56.1% #143 /229
- MARS-Bench 30.4% #14 /15
- BigCodeBench-Hard 32.5% #9 /96
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 | |
|---|---|---|---|---|
| Vectara Hallucination Leaderboard (HHEM) | 2.90% | #42 of 148 | Official leaderboard | |
| MMLU-Pro | 76.1% | #70 of 212 | Official leaderboard | |
| Fiction.LiveBench | 66.7% | #21 of 57 | Fiction.live leaderboard, via Epoch AI | |
| SuperGPQA | 50.1% | #22 of 59 | Official leaderboard | |
| LMArena (Elo) | 1,374 | #159 of 335 | Official leaderboard | |
| LLM Confabulation Leaderboard | 21.8% | #42 of 58 | Official leaderboard | |
| Phare | 62.2% | #45 of 62 | Official leaderboard | |
| MARS-Bench | 30.4% | #14 of 15 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| GSM8K Platinum Initial 268 | 99.3% | #7 of 19 | Official leaderboard | |
| MATH Level 5 | 67.2% | #36 of 94 | Epoch AI run | |
| LiveMathBench | 42.2% | #3 of 5 | Official leaderboard | |
| GPQA Diamond | 56.1% | #143 of 229 | Epoch AI run | |
| OTIS Mock AIME 2024-2025 | 16.1% | #158 of 212 | Epoch AI run | |
| Misguided Attention | 31.4% | #34 of 45 | Official leaderboard | |
| FrontierMath | 1.03% | #60 of 70 | Epoch AI run | |
| PolyMath | 19.4% | #34 of 38 | Official leaderboard | |
| PHYBench | 6.27% | #28 of 28 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| BigCodeBench-Hard | 32.5% | #9 of 96 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Step Race | 1 | #38 of 59 | Official leaderboard | |
| BrowseComp-ZH | 7.60% | #7 of 9 | Official leaderboard | |
| Elimination Game | 2 | #51 of 56 | Official leaderboard |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| FinanceReasoning | 86% | #2 of 6 | Official leaderboard |
Qwen's models over time
Qwen's models in color, against every other model on the Epoch Capabilities Index, by release date.
Source: Epoch Capabilities Index, Epoch AI (CC BY 4.0)
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.