Reasoning benchmark
MRCR v2 64K-128K leaderboard
OpenAI MRCR v2 8-needle 64K-128K. Every model the catalog carries a published MRCR v2 64K-128K value for, ranked by that value.
CategoryReasoning
MeasureLong-context retrieval
Tasks8-needle retrieval tasks
DifficultyLong-context reasoning
Published byIntroducing GPT-5.4 mini and nano
MRCR v2 slice focused on long-context retrieval at 64K-128K lengths.
MRCR v2 64K-128K ranking
2 models with a published MRCR v2 64K-128K value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Long-context retrieval |
|---|---|---|---|
| 1 |  Gemini 3.7 Flash | 97 | |
| 2 |  GPT-5.5 | OpenAI | 83.1 |
Evidence key: Observed
Rows are ordered by the value Introducing GPT-5.4 mini and nano published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.