Benchmark
Vals EMB v1.1
Scores are percentages.
Benchmark sourceScores from Vals AI.
- Models measured
- 44
- Scores recorded
- 44
- Sources
- 1
- IQ weight
- 2.89
Score distribution
- 68.3%
- Mean
- 69.6%
- Median
- 7.9%
- Std dev
- 59.8%
- 10th percentile
- 75.9%
- 90th percentile
- 40.8%
- Min
This benchmark has a weight of 2.89 for chat models, measured over 44 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.