QVQ-72B-Preview
QVQ-72B-Preview places in the top three on MME-CoT.
13
Benchmarks with a score
0
Ranked first, out of five or more models
1
In the top three
Qwen Dec 2024 QVQ-72B-Preview
Open weights Flagship
No Epoch Capabilities Index score yet
- VisualPuzzles 37.8% #14 /29
- HumanEval-V 19% #7 /24
- MM-HELIX 17.7% #10 /17
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.
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| U-MATH | 65% | #7 of 30 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| HumanEval-V | 19% | #7 of 24 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ScienceBoard | 0.49% | #15 of 15 | Official leaderboard |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MedXpertQA-MM | 33.5% | #5 of 9 | Official leaderboard | |
| MedXpertQA | 27.1% | #3 of 4 | Official leaderboard |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MathVista | 71.4% | #7 of 43 | Official leaderboard | |
| MME-CoT | 62% | #2 of 10 | Official leaderboard | |
| MMMU | 70.3% | #32 of 136 | Official leaderboard | |
| MME-Reasoning | 35.2% | #5 of 11 | Official leaderboard | |
| VisualPuzzles | 37.8% | #14 of 29 | Official leaderboard | |
| SpatialViz-Bench | 31.3% | #8 of 14 | Official leaderboard | |
| MM-HELIX | 17.7% | #10 of 17 | Official leaderboard | |
| MATH-Vision (MathVision) | 35.9% | #67 of 105 | Official leaderboard |