Benchmark
Video-MME-v2 - Capability Consistency
Vision benchmarks measure how well models understand images. Scores are percentages.
- Models measured
- 20
- Scores recorded
- 78
- Sources
- 1
- IQ weight
- 1.54
Score distribution
- 23%
- Mean
- 20.9%
- Median
- 9.1%
- Std dev
- 13.3%
- 10th percentile
- 37.1%
- 90th percentile
- 8.5%
- Min
This benchmark has a weight of 1.54 for chat models, measured over 20 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 51 | Qwen3.5 397B A17Bwo sub @ 64, instruct· Video-MME-v2 | 18.2% |
| 52 | Qwen3 VL 235Bwo sub @ 64, instruct· Video-MME-v2 | 17.7% |
| 53 | Qwen3.5 Omni Pluswo sub @ 64· Video-MME-v2 | 17.3% |
| 54 | Qwen3 VL 235Bwo sub @ 64, think· Video-MME-v2 | 17.2% |
| 55 | internvl3-5-241b-a28bwo sub @ 64, instruct· Video-MME-v2 | 17.2% |
| 56 | Qwen: Qwen3 VL 8B Thinkingw. sub @ 64, instruct· Video-MME-v2 | 17.1% |
| 57 | Qwen3.5 122B-A10Bwo sub @ 64, think· Video-MME-v2 | 16.8% |
| 58 | Qwen3.5 9Bw. sub @ 64, instruct· Video-MME-v2 | 16.7% |
| 59 | Qwen3.5 35B A3Bwo sub @ 64, think· Video-MME-v2 | 16.1% |
| 60 | qwen3-vl-30bw. sub @ 64, instruct· Video-MME-v2 | 15.7% |
| 61 | Qwen: Qwen3 VL 8B Thinkingwo sub @ 512, think· Video-MME-v2 | 15.1% |
| 62 | qwen3-vl-30bwo sub @ 512, think· Video-MME-v2 | 14.8% |
| 63 | Qwen3.5 27Bwo sub @ 64, instruct· Video-MME-v2 | 14.8% |
| 64 | GLM 4.5Vwo sub @ 64, think· Video-MME-v2 | 14.7% |
| 65 | Qwen3.5 4Bw. sub @ 64, instruct· Video-MME-v2 | 14.6% |
| 66 | Qwen3.5 122B-A10Bwo sub @ 512, think· Video-MME-v2 | 14.6% |
| 67 | qwen3-vl-30bwo sub @ 64, think· Video-MME-v2 | 14.1% |
| 68 | Qwen2.5 VL 72B Instructwo sub @ 64, instruct· Video-MME-v2 | 14.1% |
| 69 | Qwen3.5 9Bwo sub @ 64, think· Video-MME-v2 | 13.9% |
| 70 | Qwen3.5 122B-A10Bwo sub @ 64, instruct· Video-MME-v2 | 13.3% |
| 71 | Qwen: Qwen3 VL 8B Thinkingwo sub @ 64, instruct· Video-MME-v2 | 13.2% |
| 72 | Qwen3.5 35B A3Bwo sub @ 512, think· Video-MME-v2 | 12.8% |
| 73 | Qwen: Qwen3 VL 8B Thinkingwo sub @ 64, think· Video-MME-v2 | 12.7% |
| 74 | Qwen3.5 4Bwo sub @ 64, think· Video-MME-v2 | 12.4% |
| 75 | Qwen3.5 35B A3Bwo sub @ 64, instruct· Video-MME-v2 | 12.4% |
| 76 | Qwen3.5 9Bwo sub @ 64, instruct· Video-MME-v2 | 10.6% |
| 77 | Qwen3.5 4Bwo sub @ 64, instruct· Video-MME-v2 | 9.8% |
| 78 | qwen3-vl-30bwo sub @ 64, instruct· Video-MME-v2 | 8.5% |
Showing 51–78 of 78 results.