Benchmark
MIEB(lite)
Vision benchmarks measure how well models understand images. Scores are percentages.
- Models measured
- 4
- Scores recorded
- 4
- Sources
- 1
- IQ weight
- No weight published
Score distribution
- 53.9%
- Mean
- 61.1%
- Median
- 8.9%
- Std dev
- 41.6%
- 10th percentile
- 63.5%
- 90th percentile
- 41.6%
- Min