Qwen2.5-VL-32B
Qwen2.5-VL-32B ranks first on 2 of the 39 benchmarks with five or more models scored.
40
Benchmarks with a score
2
Ranked first, out of five or more models
4
In the top three
Qwen2.5-VL-32B
Open weights Mid-size
No Epoch Capabilities Index score yet
- MM-BrowseComp 1.00% #10 /11
- HiPhO 28% #31 /36
- Video SimpleQA 30.7% #10 /27
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 | |
|---|---|---|---|---|
| MMBench-CN v1.1 | 83.2% | #27 of 178 | Official leaderboard | |
| Kaleidoscope | 48.2% | #2 of 7 | Official leaderboard | |
| Video SimpleQA | 30.7% | #10 of 27 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| HiPhO | 28% | #31 of 36 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SArena | 0.84 | #7 of 17 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ScreenSpot-Pro | 48% | #5 of 11 | Official leaderboard | |
| BrowseComp-VL | 7.20% | #3 of 4 | Official leaderboard | |
| GameHorizon | 60.8 | #31 of 37 | Official leaderboard | |
| SnakeBench | 17 | #160 of 188 | Official leaderboard | |
| MM-BrowseComp | 1.00% | #10 of 11 | Official leaderboard | |
| OSWorld-Verified | 3.88% | #33 of 33 | Official leaderboard |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| BizFinBench | 68.6% | #1 of 13 | Official leaderboard |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MMStar | 70.3% | #8 of 173 | Official leaderboard | |
| MM-Vet | 72.9% | #16 of 164 | Official leaderboard | |
| MMBench-EN v1.1 | 84.7% | #20 of 178 | Official leaderboard | |
| Video-Holmes | 38.4% | #1 of 5 | Official leaderboard | |
| OCR-Reasoning | 36.2% | #4 of 15 | Official leaderboard | |
| CrossVid | 33.7% | #4 of 13 | Official leaderboard | |
| SpatialViz-Bench | 33.9% | #5 of 14 | Official leaderboard | |
| OmniVideoBench | 31.8% | #3 of 8 | Official leaderboard | |
| MM-HELIX | 20.6% | #7 of 17 | Official leaderboard | |
| Point-Bench | 59% | #7 of 17 | Official leaderboard | |
| VHELM | 58.6 | #18 of 41 | Official leaderboard | |
| MORSE-500 | 15.6% | #5 of 11 | Official leaderboard | |
| EmoBench-M | 56.7% | #7 of 15 | Official leaderboard | |
| MME-Reasoning | 33.2% | #8 of 11 | Official leaderboard | |
| MeasureBench Real-world | 11.7% | #14 of 19 | Official leaderboard | |
| MeasureBench Synthetic | 10.5% | #14 of 19 | Official leaderboard | |
| RefAdv-S | 50.6% | #14 of 19 | Official leaderboard | |
| GSM8K-V | 21.8% | #18 of 24 | Official leaderboard | |
| MathCanvas-Bench | 15.4% | #12 of 16 | Official leaderboard | |
| MME-CC | 14.4% | #9 of 12 | Official leaderboard | |
| RoboRewardBench | 1.14 | #17 of 22 | Official leaderboard | |
| Math-VR | 10% | #21 of 26 | Official leaderboard | |
| Vision Arena | 1,118 | #112 of 136 | Official leaderboard | |
| SpaCE-10 | 32.6% | #31 of 36 | Official leaderboard | |
| Progress-Bench | 40.8 | #8 of 9 | Official leaderboard | |
| VisFactor | 13.3% | #25 of 26 | Official leaderboard |
Safety and security
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Video-SafetyBench | 2.00 | #7 of 15 | Official leaderboard |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| OpenVLM Leaderboard | 74.8 | #14 of 173 | Official leaderboard |