Skip to content

Benchmark

MTEB(pol, v1)

Embedding benchmarks measure how well models find related items, for example when matching a search to relevant results. Scores are percentages.

Models measured
17
Scores recorded
17
Sources
1
IQ weight
1.34

Score distribution

56.4%
Mean
51.2%
Median
10%
Std dev
47.3%
10th percentile
69.9%
90th percentile
44.2%
Min
44.2%73.5%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.34 for embedding models, measured over 17 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard