DeepSeek-VL2-Small
DeepSeek-VL2-Small ranks first on MindCube.
18
Benchmarks with a score
1
Ranked first, out of five or more models
1
In the top three
DeepSeek Dec 2024 DeepSeek-VL2-Small
Open weights Small
No Epoch Capabilities Index score yet
- Video SimpleQA 7.40% #26 /27
- MindCube 47.6% #1 /7
- MM-Vet 58.6% #60 /164
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 | 79.1% | #57 of 178 | Official leaderboard | |
| Video SimpleQA | 7.40% | #26 of 27 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| PHYSICS | 1.70% | #20 of 20 | Official leaderboard | |
| PHYSICS-Verified | 3.00% | #19 of 19 | Official leaderboard |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MindCube | 47.6% | #1 of 7 | Official leaderboard | |
| CoreCognition | 61.4% | #19 of 119 | Official leaderboard | |
| MMBench | 82.8% | #34 of 155 | Official leaderboard | |
| MMBench-EN v1.1 | 80.7% | #49 of 178 | Official leaderboard | |
| CCBench (MMBench Chinese Culture) | 56.7% | #45 of 152 | Official leaderboard | |
| MM-Vet | 58.6% | #60 of 164 | Official leaderboard | |
| MMStar | 57.5% | #67 of 173 | Official leaderboard | |
| MMSI-Bench | 28.6% | #12 of 28 | Official leaderboard | |
| LEGO-Puzzles | 37.5% | #7 of 16 | Official leaderboard | |
| MMVU | 46.9% | #15 of 28 | Official leaderboard | |
| MMMU | 48% | #93 of 136 | Official leaderboard | |
| KRETA | 53.3% | #12 of 17 | Official leaderboard | |
| All-Angles Bench | 45.5% | #13 of 18 | Official leaderboard |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| OpenVLM Leaderboard | 64.5 | #60 of 173 | Official leaderboard |