MRCR 128K (4-needle)
reasoning official site →
MRCR (Multi-Round Coreference Resolution) at 128K context length with 4 needles. Models must navigate long conversations to reproduce specific model outputs, testing attention and reasoning across 128K-token contexts with 4 items to retrieve.
Methodology
Imported from llm-stats public benchmark metadata. Modality: text. Max score: 1. Categories: general, long_context, reasoning. Language: en. Verified by llm-stats: no.