Benchmark
BLINK - Jigsaw
Vision benchmarks measure how well models understand images. Scores are percentages.
All vision benchmarksScores from EvalAI data.
- Models measured
- 4
- Scores recorded
- 4
- Sources
- 1
- IQ weight
- No weight published
Score distribution
- 36.7%
- Mean
- 54%
- Median
- 21.4%
- Std dev
- 3.3%
- 10th percentile
- 56.7%
- 90th percentile
- 3.3%
- Min
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | LLaVA 1.6 34BJigsaw· EvalAI data | 56.7% |
| 2 | Gemini ProJigsaw· EvalAI data | 54% |
| 3 | Claude 3 OpusJigsaw· EvalAI data | 32.7% |
| 4 | Qwen-VL MaxJigsaw· EvalAI data | 3.3% |