Beta version: *Information might not be fully accurate. Please report any discrepancies.

MRCR v2

Name: MRCR v2 leaderboard dataset
Creator: LLM Registry

mrcr-v2Long ContextHigher is better

Multi-Round Context Retrieval - 8-needle test.

Loading chart…

Loading leaderboard…

Benchmark Info

Current SOTA

Median Score67.2%

DistributionBalanced

Score Distribution

P10

P50

P90

0100%

LongBench v2

longbench-v2

AA-LCR

lcr

Graphwalks Bfs

graphwalks-bfs