Benchmark
MRCR 1M
MRCR 1M is a variant of the Multi-Round Coreference Resolution benchmark designed for testing extremely long context capabilities with approximately 1…
- Modality
- text
- Categories
- long_context, reasoning, general
- Openness
- unknown
- Source
- llm_stats
- Reported scores
- 4
Reported scores
llm_stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| DeepSeek-V4-Flash-0423 | DeepSeek | 0.769 | 2026-04-23 | evidence |
| DeepSeek-V4-Flash-Max | DeepSeek | 0.787 | 2026-04-23 | evidence |
| DeepSeek-V4-Pro-Max | DeepSeek | 0.835 | 2026-04-23 | evidence |
| Gemini 2.0 Flash-Lite | 0.58 | 2025-02-05 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.