Skip to content

Benchmark

MRCR v2 (8-needle, 512K-1M)

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

Models measured
4
Scores recorded
4
Sources
1
IQ weight
No weight published

Score distribution

71.4%
Mean
73.8%
Median
20.2%
Std dev
41.3%
10th percentile
98.1%
90th percentile
41.3%
Min
41.3%98.1%
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1Muse Spark 1.3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim98.1%
2GPT-5.6 Sol· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim73.8%
3GPT-5.6 Terra· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim72.5%
4GPT-5.6 Luna· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim41.3%

Data credits: Data by LLM Stats