---
title: "文档理解排行榜"
description: "文档分析与长内容推理的排名。"
canonical: https://pirouter.ai/zh/rankings/document
last-updated: 2026-10-15
---

# 文档理解排行榜

> 文档分析与长内容推理的排名。

- 更新于: 2026-09-13
- 537,459 票 · 38 个模型
- 分数 · 越高越好

## 完整排名

| Rank | Rank spread | Model | Vendor | Score | Votes |
|---|---|---|---|---|---|
| 1 | 1–7 | Claude Fable 5 | Anthropic | 1499 ±8 | 8497 |
| 2 | 1–8 | Claude Opus 4.6 (high) | Anthropic | 1495 ±7 | 27929 |
| 3 | 1–9 | Claude Opus 4.7 | Anthropic | 1494 ±6 | 22192 |
| 4 | 1–15 | [Claude Fable 5.1 \(max\)](https://pirouter.ai/zh/models/anthropic/claude-fable-5.1) | Anthropic | 1493 ±15 | 1403 |
| 5 | 1–15 | Claude Opus 5 (high) | Anthropic | 1490 ±7 | 8794 |
| 6 | 1–15 | [GPT-5.6 Sol \(xhigh\)](https://pirouter.ai/zh/models/openai/gpt-5.6-sol) | OpenAI | 1489 ±9 | 4818 |
| 7 | 1–15 | Claude Opus 4.8 (high) | Anthropic | 1486 ±7 | 11551 |
| 8 | 2–15 | [GPT-5.5](https://pirouter.ai/zh/models/openai/gpt-5.5) | OpenAI | 1483 ±6 | 21279 |
| 9 | 3–19 | Muse Spark 1.1 | Meta | 1480 ±9 | 4885 |
| 10 | 4–19 | Claude Sonnet 4.6 | Anthropic | 1478 ±6 | 57813 |
| 11 | 4–22 | [GPT-5.6 Terra \(xhigh\)](https://pirouter.ai/zh/models/openai/gpt-5.6-terra) | OpenAI | 1477 ±8 | 5787 |
| 12 | 4–22 | [Claude Sonnet 5 \(high\)](https://pirouter.ai/zh/models/anthropic/claude-sonnet-5) | Anthropic | 1476 ±8 | 7411 |
| 13 | 4–26 | [GPT-6 Astra \(max\)](https://pirouter.ai/zh/models/openai/gpt-6-astra) | OpenAI | 1473 ±14 | 1621 |
| 14 | 9–24 | [GPT-5.4](https://pirouter.ai/zh/models/openai/gpt-5.4) | OpenAI | 1470 ±6 | 33331 |
| 15 | 4–29 | Muse Spark 1.3 (max) | Meta | 1468 ±18 | 1006 |
| 16 | 4–29 | Muse Spark | Meta | 1467 ±18 | 1084 |
| 17 | 9–28 | Claude Opus 4.5 | Anthropic | 1465 ±10 | 7981 |
| 18 | 9–27 | Grok 4.5 | xAI | 1464 ±9 | 4965 |
| 19 | 11–28 | [GPT-5.6 Luna \(xhigh\)](https://pirouter.ai/zh/models/openai/gpt-5.6-luna) | OpenAI | 1462 ±8 | 5766 |
| 20 | 9–29 | Grok 4.6 (high) | xAI | 1461 ±12 | 2258 |
| 21 | 11–28 | [Gemini 3.5 Flash \(medium\)](https://pirouter.ai/zh/models/google/gemini-3.5-flash) | Google | 1461 ±8 | 6500 |
| 22 | 11–29 | GPT-5.5 Instant | OpenAI | 1460 ±9 | 8497 |
| 23 | 13–28 | Gemini 3.1 Pro Preview | Google | 1459 ±5 | 49457 |
| 24 | 13–32 | [Gemini 3.6 Flash \(high\)](https://pirouter.ai/zh/models/google/gemini-3.6-flash) | Google | 1453 ±11 | 2975 |
| 25 | 14–31 | Qwen3.7 Plus | Alibaba | 1453 ±9 | 4591 |
| 26 | 14–32 | Gemini 3 Pro | Google | 1451 ±9 | 10769 |
| 27 | 15–32 | Kimi K2.6 | Moonshot | 1450 ±8 | 11291 |
| 28 | 16–31 | Claude Sonnet 4.5 | Anthropic | 1449 ±6 | 31455 |
| 29 | 20–34 | Gemma 4 31b | Google | 1445 ±8 | 12326 |
| 30 | 24–34 | Grok 4.20 Beta 0309 Reasoning | xAI | 1439 ±7 | 20948 |
| 31 | 24–35 | Gemini 3 Flash | Google | 1439 ±9 | 7158 |
| 32 | 26–35 | Kimi K2.5 (thinking) | Moonshot | 1436 ±7 | 21569 |
| 33 | 29–37 | Minimax M3 | MiniMax | 1433 ±8 | 6314 |
| 34 | 29–37 | Gemini 2.5 Pro | Google | 1431 ±6 | 25110 |
| 35 | 31–37 | GPT-5.2 | OpenAI | 1425 ±6 | 28212 |
| 36 | 33–37 | Claude Haiku 4.5 | Anthropic | 1419 ±6 | 34677 |
| 37 | 33–38 | GPT-5.1 | OpenAI | 1416 ±9 | 8244 |
| 38 | 37–38 | GLM 5v Turbo | Z.ai | 1400 ±8 | 6995 |

## FAQ

### 文档理解最好的 AI 模型是哪个？

截至 2026-09-13，文档理解榜第一为 Anthropic Claude Fable 5，分数 1499，名次区间 1–7。

### 文档理解最好的开放权重模型是哪个？

Kimi K2.6（Moonshot），全榜第 27 名。

### 这个榜单排的是什么？

文档分析与长内容推理的排名。

### 多久更新一次？

有新模型发布或排名变化时更新；本榜最近一次更新为 2026-09-13。

数据：[LMArena leaderboard dataset](https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset)，按 [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) 许可使用；仅做筛选与推理档位合并，分数未改动。
查看原榜单: https://arena.ai/leaderboard/document