# LongBench v2 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://modelscale.dev/leaderboards/benchmarks/reasoning-long-bench-v2

Every model the catalog carries a published LongBench v2 value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | Reasoning |
| Measure | Extended-context retrieval and reasoning |
| Tasks | Long-context tasks |
| Difficulty | Hard long-context |
| Published by | [LongBench v2](https://arxiv.org/abs/2412.15204) |

A long-context benchmark that measures whether models can actually use extended context windows for reasoning and retrieval.

## LongBench v2 ranking

26 models with a published LongBench v2 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 | Extended-context retrieval and reasoning | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [Inkling 2 Small](https://modelscale.dev/models/inkling-2-small) | Thinking Machines Lab | 70.6 | Observed 2026-11-21 · source benchlm:models |
| 2 | [Beam 1.0 501B-A23B FP8](https://modelscale.dev/models/reflection-beam-fp8) | Reflection AI | 67.6 | Observed 2026-11-21 · source benchlm:models |
| 3 | [Beam 1.0 501B-A23B](https://modelscale.dev/models/reflection-beam) | Reflection AI | 67.5 | Observed 2026-11-21 · source benchlm:models |
| 4 | [Beam 1.0 501B-A23B NVFP4](https://modelscale.dev/models/reflection-beam-nvfp4) | Reflection AI | 67.2 | Observed 2026-11-21 · source benchlm:models |
| 5 | [Inkling 2 Mini](https://modelscale.dev/models/inkling-2-mini) | Thinking Machines Lab | 66.9 | Observed 2026-11-21 · source benchlm:models |
| 6 | [Qwen3.8 Max](https://modelscale.dev/models/qwen3-8-max) | Alibaba | 66.3 | Observed 2026-11-21 · source benchlm:models |
| 7 | [Claude Opus 4.5](https://modelscale.dev/models/claude-opus-4-5) | Anthropic | 64.4 | Observed 2026-11-21 · source benchlm:models |
| 8 | [Qwen3.5 397B](https://modelscale.dev/models/qwen3-5-397b) | Alibaba | 63.2 | Observed 2026-11-21 · source benchlm:models |
| 9 | [Qwen3.6 Plus](https://modelscale.dev/models/qwen3-6-plus) | Alibaba | 62 | Observed 2026-11-21 · source benchlm:models |
| 10 | [Nemotron 3 Ultra](https://modelscale.dev/models/nemotron-3-ultra-500b) | NVIDIA | 61.9 | Observed 2026-11-21 · source benchlm:models |
| 11 | [MiniMax M1 80k](https://modelscale.dev/models/minimax-m1-80k) | MiniMax | 61.5 | Observed 2026-11-21 · source benchlm:models |
| 12 | [Kimi K2.5](https://modelscale.dev/models/kimi-k2-5) | Moonshot AI | 61 | Observed 2026-11-21 · source benchlm:models |
| 12 | [Kimi K2.5 (Reasoning)](https://modelscale.dev/models/kimi-k2-5-reasoning) | Moonshot AI | 61 | Last good 2026-11-18 · source benchlm:models |
| 14 | [GLM-5](https://modelscale.dev/models/glm-5) | Z.AI | 60.8 | Observed 2026-11-21 · source benchlm:models |
| 15 | [MiMo-V2-Flash](https://modelscale.dev/models/mimo-v2-flash) | Xiaomi | 60.6 | Observed 2026-11-21 · source benchlm:models |
| 15 | [Qwen3.5-27B](https://modelscale.dev/models/qwen3-5-27b) | Alibaba | 60.6 | Observed 2026-11-21 · source benchlm:models |
| 17 | [Agents-A1](https://modelscale.dev/models/agents-a1) | InternScience | 60.2 | Observed 2026-11-21 · source benchlm:models |
| 17 | [Qwen3.5-122B-A10B](https://modelscale.dev/models/qwen3-5-122b-a10b) | Alibaba | 60.2 | Observed 2026-11-21 · source benchlm:models |
| 19 | [Qwen3.5-35B-A3B](https://modelscale.dev/models/qwen3-5-35b-a3b) | Alibaba | 59 | Observed 2026-11-21 · source benchlm:models |
| 20 | [Spark-X2.6-4B](https://modelscale.dev/models/spark-x2-6-4b) | SparkLLM | 53.3 | Observed 2026-11-21 · source benchlm:models |
| 21 | [Agents-A1-4B](https://modelscale.dev/models/agents-a1-4b) | InternScience | 52.1 | Observed 2026-11-21 · source benchlm:models |
| 22 | [Nemotron 3.5 Bolt E4B BF16](https://modelscale.dev/models/nemotron-3-5-bolt-e4b-bf16) | NVIDIA | 50.5 | Observed 2026-11-21 · source benchlm:models |
| 23 | [DeepSeek V3](https://modelscale.dev/models/deepseek-v3) | DeepSeek | 48.7 | Observed 2026-11-21 · source benchlm:models |
| 24 | [Nemotron 3.5 Bolt E4B](https://modelscale.dev/models/nemotron-3-5-bolt-e4b) | NVIDIA | 48.67 | Observed 2026-11-21 · source benchlm:models |
| 25 | [MiniCPM5-2B](https://modelscale.dev/models/minicpm5-2b) | OpenBMB | 43.7 | Observed 2026-11-21 · source benchlm:models |
| 26 | [LLaDA2.2-mini](https://modelscale.dev/models/llada2-2-mini) | InclusionAI | 34.99 | 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.
