Skip to content

Benchmark

GDM-MRCRv2: 8 needle 32k

Long context benchmarks measure how well models use information in long inputs. Scores are percentages.

Models measured
68
Scores recorded
177
Sources
1
IQ weight
1.85

Score distribution

60.7%
Mean
65.5%
Median
27.2%
Std dev
23.8%
10th percentile
94.2%
90th percentile
11.8%
Min
11.8%99.3%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.85 for chat models, measured over 68 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

Showing 151–177 of 177 results.