Skip to content

Benchmark

SuperCLUE 超长上下文

Scores are percentages.

Models measured
20
Scores recorded
20
Sources
1
IQ weight
1.83

Score distribution

56.5%
Mean
49.4%
Median
21.2%
Std dev
35.7%
10th percentile
88.1%
90th percentile
30.2%
Min
30.2%95.6%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.83 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