# MRCRv2 leaderboard

Reasoning benchmark

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelchoice.dev/leaderboards/benchmarks/reasoning-mrcrv2

Every model the catalog carries a published MRCRv2 value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | Reasoning |
| Measure | Multi-round long-context evaluation |
| Tasks | Long-context retrieval |
| Difficulty | Hard long-context |
| Published by | [Introducing GPT-5.2 and GPT-5.2 Pro](https://openai.com/index/introducing-gpt-5-2/) |

A long-context benchmark for memory, retrieval, and multi-round coherence over large contexts.

## MRCRv2 ranking

14 models with a published MRCRv2 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 | Multi-round long-context evaluation | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [Sakana Fugu-Ultra](https://modelchoice.dev/models/sakana-fugu-ultra) | Sakana AI | 93.6 | Observed 2026-11-21 · source benchlm:models |
| 2 | [Qwen3.8 Max](https://modelchoice.dev/models/qwen3-8-max) | Alibaba | 92.9 | Observed 2026-11-21 · source benchlm:models |
| 3 | [Qwen3.7 Plus](https://modelchoice.dev/models/qwen3-7-plus) | Alibaba | 91.7 | Observed 2026-11-21 · source benchlm:models |
| 4 | [Qwen3.7 Max](https://modelchoice.dev/models/qwen3-7-max) | Alibaba | 90.4 | Observed 2026-11-21 · source benchlm:models |
| 5 | [Sakana Fugu](https://modelchoice.dev/models/sakana-fugu) | Sakana AI | 86.6 | Observed 2026-11-21 · source benchlm:models |
| 6 | [Gemini 3.5 Flash](https://modelchoice.dev/models/gemini-3-5-flash) | Google | 77.3 | Observed 2026-11-21 · source benchlm:models |
| 7 | [Mercury 3](https://modelchoice.dev/models/mercury-3) | Inception | 75.5 | Observed 2026-11-21 · source benchlm:models |
| 8 | [Gemini 3.5 Flash-Lite](https://modelchoice.dev/models/gemini-3-5-flash-lite) | Google | 72.2 | Observed 2026-11-21 · source benchlm:models |
| 9 | [Gemma 4 31B](https://modelchoice.dev/models/gemma-4-31b) | Google | 66.4 | Observed 2026-11-21 · source benchlm:models |
| 10 | [Pokee-Isaac 28B](https://modelchoice.dev/models/pokee-isaac-28b) | Pokee AI | 60.7 | Observed 2026-11-21 · source benchlm:models |
| 11 | [Gemma 4 26B A4B](https://modelchoice.dev/models/gemma-4-26b-a4b) | Google | 44.1 | Observed 2026-11-21 · source benchlm:models |
| 12 | [Gemma 4 12B](https://modelchoice.dev/models/gemma-4-12b) | Google | 43.4 | Observed 2026-11-21 · source benchlm:models |
| 13 | [Gemma 4 E4B](https://modelchoice.dev/models/gemma-4-e4b) | Google | 25.4 | Observed 2026-11-21 · source benchlm:models |
| 14 | [Gemma 4 E2B](https://modelchoice.dev/models/gemma-4-e2b) | Google | 19.1 | Observed 2026-11-21 · source benchlm:models |

Rows are ordered by the value the source published, highest first. The source does not state whether a higher value is the better result, so this document does not either: for a benchmark that measures a rate of failure, read the table from the bottom.
