1
A
claude-opus-4-6-thinking
claude-opus-4-6-thinkingAnthropic
100.0
11.9K
1M
¥36 / ¥180Input/Output
Chat · Document · Overall Leaderboard
Ranking for Document / Overall, based on public preference data.
Overall model ranking guide
Ranking for Document / Overall, based on public preference data.
claude-opus-4-6-thinkingclaude-opus-4-6claude-opus-4-7claude-opus-4-7-thinkinggpt-5.5-high
2
Aclaude-opus-4-6
claude-opus-4-6Anthropic
95.7
20.4K
1M
¥36 / ¥180Input/Output
3
Aclaude-opus-4-7
claude-opus-4-7Anthropic
91.3
6.7K
1M
¥36 / ¥180Input/Output
4
Aclaude-opus-4-7-thinking
claude-opus-4-7-thinkingAnthropic
87.0
6.4K
1M
¥36 / ¥180Input/Output
5
O
gpt-5.5-high
gpt-5.5-highOpenai
82.6
4.6K
1.05M
¥36 / ¥216Input/Output
6
Aclaude-sonnet-4-6
claude-sonnet-4-6Anthropic
78.3
31.9K
1M
¥21.6 / ¥108Input/Output
7
Ogpt-5.5
gpt-5.5Openai
73.9
4.7K
1.05M
¥36 / ¥216Input/Output
8
Ogpt-5.4
gpt-5.4Openai
69.6
14.4K
1.05M
¥18 / ¥108Input/Output
9
Aclaude-opus-4-5-20251101
claude-opus-4-5-20251101Anthropic
65.2
8K
200K
¥36 / ¥180Input/Output
10
M
kimi-k2.6
kimi-k2.6Moonshot
60.9
3.8K
262K
¥6.84 / ¥28.8Input/Output
11
M
muse-spark
muse-sparkMeta
56.5
868
-
-
12
Aclaude-sonnet-4-5-20250929
claude-sonnet-4-5-20250929Anthropic
52.2
16.7K
200K
¥21.6 / ¥108Input/Output
13
G
gemini-3.1-pro-preview
gemini-3.1-pro-previewGoogle
47.8
24.9K
1.05M
¥14.4 / ¥86.4Input/Output
14
Ggemini-3-pro
gemini-3-proGoogle
43.5
10.8K
1.05M
¥14.4 / ¥86.4Input/Output
15
Mkimi-k2.5-thinking
kimi-k2.5-thinkingMoonshot
39.1
10.5K
262K
¥4.32 / ¥21.6Input/Output
16
Ggemini-2.5-pro
gemini-2.5-proGoogle
34.8
20K
1.05M
¥9 / ¥72Input/Output
17
Ggemma-4-31b
gemma-4-31bGoogle
30.4
4.4K
262K
¥3.24 / ¥7.2Input/Output
18
Aclaude-haiku-4-5-20251001
claude-haiku-4-5-20251001Anthropic
26.1
17.9K
200K
¥7.2 / ¥36Input/Output
19
X
grok-4.20-beta-0309-reasoning
grok-4.20-beta-0309-reasoningXai
21.7
6.8K
2M
¥14.4 / ¥43.2Input/Output
20
Ggemini-3-flash
gemini-3-flashGoogle
17.4
7.2K
1.05M
¥3.6 / ¥21.6Input/Output
21
Ogpt-5.2-high
gpt-5.2-highOpenai
13.0
7.1K
400K
¥12.6 / ¥101Input/Output
22
Ogpt-5.2
gpt-5.2Openai
8.7
22.4K
400K
¥12.6 / ¥101Input/Output
23
Ogpt-5.5-instant
gpt-5.5-instantOpenai
4.3
3.5K
400K
¥9 / ¥72Input/Output
24
Ogpt-5.1
gpt-5.1Openai
0.0
8.3K
400K
¥9 / ¥72Input/Output
claude-opus-4-6-thinking why it ranks first
claude-opus-4-6-thinking ranks first with a percent score of 100.0 and 11.9K samples. Use it as the first option for this leaderboard, then compare price, context and availability.
Do not only look at rank #1
Start with the leaderboard closest to your task. Compare the top models by score and sample size, then check price, context length, open or closed access, and provider availability.
Compare adjacent capabilities
FAQ
文档理解总榜排行榜看什么指标？
主要看排名、百分制分数、样本量和来源。分数用于快速比较同一榜单内模型表现，样本量用于判断结果稳定性。
为什么不同榜单不能直接混合成总分？
不同榜单的任务、样本和评测口径不同，模力榜默认只在同一榜单内排序，避免把写作、代码、图像等能力强行合并。
文档理解总榜模型应该怎么选？
优先看与你任务最接近的榜单，再结合价格、上下文长度、开源闭源和厂商可用性。排名靠前不代表适合所有预算和部署方式。
榜单多久更新？
页面展示的是最新成功采集的公开榜单数据。当前优先使用 LMArena leaderboard dataset，并在页面来源中保留原始链接。