1
A
claude-opus-4-7
claude-opus-4-7Anthropic
100.0
4.8K
1M
¥36 / ¥180Input/Output
Chat · Vision · OCR Leaderboard
Ranking for Vision / OCR, based on public preference data.
OCR model ranking guide
Ranking for Vision / OCR, based on public preference data.
claude-opus-4-7claude-opus-4-6-thinkingclaude-opus-4-7-thinkingclaude-opus-4-6gemini-3-pro
2
Aclaude-opus-4-6-thinking
claude-opus-4-6-thinkingAnthropic
98.8
4.9K
1M
¥36 / ¥180Input/Output
3
Aclaude-opus-4-7-thinking
claude-opus-4-7-thinkingAnthropic
97.5
4.5K
1M
¥36 / ¥180Input/Output
4
Aclaude-opus-4-6
claude-opus-4-6Anthropic
96.3
6.1K
1M
¥36 / ¥180Input/Output
5
G
gemini-3-pro
gemini-3-proGoogle
95.1
8.1K
1.05M
¥14.4 / ¥86.4Input/Output
6
M
muse-spark
muse-sparkMeta
93.8
3.2K
-
-
7
Ggemini-3.1-pro-preview
gemini-3.1-pro-previewGoogle
92.6
11.4K
1.05M
¥14.4 / ¥86.4Input/Output
8
O
gpt-5.4-high
gpt-5.4-highOpenai
91.4
4.2K
1.05M
¥18 / ¥108Input/Output
9
Ogpt-5.5
gpt-5.5Openai
90.1
3.3K
1.05M
¥36 / ¥216Input/Output
10
Ogpt-5.5-high
gpt-5.5-highOpenai
88.9
3K
1.05M
¥36 / ¥216Input/Output
11
Aclaude-sonnet-4-6
claude-sonnet-4-6Anthropic
87.7
6.3K
1M
¥21.6 / ¥108Input/Output
12
Ogpt-5.4
gpt-5.4Openai
86.4
3.9K
1.05M
¥18 / ¥108Input/Output
13
Ggemini-3-flash
gemini-3-flashGoogle
85.2
14.2K
1.05M
¥3.6 / ¥21.6Input/Output
14
M
kimi-k2.6
kimi-k2.6Moonshot
84.0
4K
262K
¥6.84 / ¥28.8Input/Output
15
B
dola-seed-2.0-pro
dola-seed-2.0-proBytedance
82.7
6.2K
-
-
16
A
qwen3.7-plus-preview
qwen3.7-plus-previewAlibaba
81.5
2.7K
131K
¥3.6 / ¥21.6Input/Output
17
Ogpt-5.2-chat-latest-20260210
gpt-5.2-chat-latest-20260210Openai
80.2
8.5K
400K
¥12.6 / ¥101Input/Output
18
Mkimi-k2.5-thinking
kimi-k2.5-thinkingMoonshot
79.0
9.7K
262K
¥4.32 / ¥21.6Input/Output
19
Aqwen3.5-397b-a17b
qwen3.5-397b-a17bAlibaba
77.8
8K
262K
¥3.1 / ¥18.6Input/Output
20
Ggemini-3-flash (thinking-minimal)
gemini-3-flash (thinking-minimal)Google
76.5
12.9K
1.05M
¥3.6 / ¥21.6Input/Output
21
Ggemini-2.5-pro
gemini-2.5-proGoogle
75.3
30.8K
1.05M
¥9 / ¥72Input/Output
22
Ggemma-4-31b
gemma-4-31bGoogle
74.1
11.9K
262K
¥3.24 / ¥7.2Input/Output
23
Ggemma-4-26b-a4b
gemma-4-26b-a4bGoogle
72.8
7.3K
262K
¥0.94 / ¥2.88Input/Output
24
Z
glm-5v-turbo
glm-5v-turboZai
71.6
5.2K
200K
¥0 / ¥0Input/Output
25
Mkimi-k2.5-instant
kimi-k2.5-instantMoonshot
70.4
2.6K
262K
¥4.32 / ¥21.6Input/Output
26
X
grok-4.20-beta-0309-reasoning
grok-4.20-beta-0309-reasoningXai
69.1
6.6K
2M
¥14.4 / ¥43.2Input/Output
27
Ggemini-2.5-flash-preview-09-2025
gemini-2.5-flash-preview-09-2025Google
67.9
2.8K
1M
¥2.16 / ¥18Input/Output
28
Ogpt-5.2-high
gpt-5.2-highOpenai
66.7
10.2K
400K
¥12.6 / ¥101Input/Output
29
Ogpt-5.4-mini-high
gpt-5.4-mini-highOpenai
65.4
5.9K
400K
¥5.4 / ¥32.4Input/Output
30
Aqwen3-vl-235b-a22b-instruct
qwen3-vl-235b-a22b-instructAlibaba
64.2
7.6K
128K
¥2.16 / ¥8.64Input/Output
31
Xgrok-4.20-multi-agent-beta-0309
grok-4.20-multi-agent-beta-0309Xai
63.0
6.1K
2M
¥14.4 / ¥43.2Input/Output
32
Ogpt-5.5-instant
gpt-5.5-instantOpenai
61.7
2.7K
400K
¥9 / ¥72Input/Output
33
MI
mimo-v2.5
mimo-v2.5Xiaomi
60.5
4.9K
1.05M
¥2.88 / ¥14.4Input/Output
34
Ogpt-5.1-high
gpt-5.1-highOpenai
59.3
5.7K
400K
¥9 / ¥72Input/Output
35
B
ernie-5.0-preview-1220
ernie-5.0-preview-1220Baidu
58.0
1.9K
128K
¥7.92 / ¥14.4Input/Output
36
Ochatgpt-4o-latest-20250326
chatgpt-4o-latest-20250326Openai
56.8
11.2K
128K
¥18 / ¥72Input/Output
37
Xgrok-4.3
grok-4.3Xai
55.6
2.8K
1M
¥9 / ¥18Input/Output
38
Ggemini-3.1-flash-lite-preview
gemini-3.1-flash-lite-previewGoogle
54.3
9.5K
1.05M
¥1.8 / ¥10.8Input/Output
39
Aqwen3.5-122b-a10b
qwen3.5-122b-a10bAlibaba
53.1
6.9K
262K
¥2.88 / ¥23Input/Output
40
Ogpt-5-chat
gpt-5-chatOpenai
51.9
10.8K
400K
¥9 / ¥72Input/Output
41
Aqwen3.5-27b
qwen3.5-27bAlibaba
50.6
6.5K
262K
¥2.16 / ¥17.3Input/Output
42
Ogpt-5.1
gpt-5.1Openai
49.4
6.5K
400K
¥9 / ¥72Input/Output
43
Ggemini-2.5-flash
gemini-2.5-flashGoogle
48.1
25K
1.05M
¥2.16 / ¥18Input/Output
44
Aqwen-vl-max-2025-08-13
qwen-vl-max-2025-08-13Alibaba
46.9
1.2K
131K
¥1.66 / ¥4.13Input/Output
45
Ogpt-5.2
gpt-5.2Openai
45.7
10.9K
400K
¥12.6 / ¥101Input/Output
46
MImimo-v2-omni
mimo-v2-omniXiaomi
44.4
5.2K
262K
¥2.88 / ¥14.4Input/Output
47
Oo3-2025-04-16
o3-2025-04-16Openai
43.2
14.7K
200K
¥14.4 / ¥57.6Input/Output
48
Ogpt-5-high
gpt-5-highOpenai
42.0
10.9K
400K
¥9 / ¥72Input/Output
49
Ogpt-4.1-2025-04-14
gpt-4.1-2025-04-14Openai
40.7
11.6K
1.05M
¥14.4 / ¥57.6Input/Output
50
Aqwen3-vl-235b-a22b-thinking
qwen3-vl-235b-a22b-thinkingAlibaba
39.5
1.4K
131K
¥2.06 / ¥8.26Input/Output
51
Ogpt-5.4-nano-high
gpt-5.4-nano-highOpenai
38.3
5.9K
400K
¥1.44 / ¥9Input/Output
52
Ggemini-2.5-flash-lite-preview-09-2025-no-thinking
gemini-2.5-flash-lite-preview-09-2025-no-thinkingGoogle
37.0
2.8K
1.05M
¥0.72 / ¥2.88Input/Output
53
Ogpt-5-mini-high
gpt-5-mini-highOpenai
35.8
8.2K
400K
¥1.8 / ¥14.4Input/Output
54
Oo4-mini-2025-04-16
o4-mini-2025-04-16Openai
34.6
12.1K
200K
¥7.92 / ¥31.7Input/Output
55
Aclaude-sonnet-4-20250514-thinking-32k
claude-sonnet-4-20250514-thinking-32kAnthropic
33.3
744
200K
¥21.6 / ¥108Input/Output
56
Aclaude-opus-4-20250514-thinking-16k
claude-opus-4-20250514-thinking-16kAnthropic
32.1
852
200K
¥108 / ¥540Input/Output
57
Xgrok-4-0709
grok-4-0709Xai
30.9
10K
256K
¥21.6 / ¥108Input/Output
58
Ogpt-4.1-mini-2025-04-14
gpt-4.1-mini-2025-04-14Openai
29.6
10.7K
1.05M
¥2.88 / ¥11.5Input/Output
59
Ggemini-2.5-flash-lite-preview-06-17-thinking
gemini-2.5-flash-lite-preview-06-17-thinkingGoogle
28.4
10.7K
65.5K
¥0.72 / ¥2.88Input/Output
60
Xgrok-4-1-fast-reasoning
grok-4-1-fast-reasoningXai
27.2
7.3K
2M
¥1.44 / ¥3.6Input/Output
61
Aclaude-3-7-sonnet-20250219-thinking-32k
claude-3-7-sonnet-20250219-thinking-32kAnthropic
25.9
887
-
-
62
TE
hunyuan-vision-1.5-thinking
hunyuan-vision-1.5-thinkingTencent
24.7
1.4K
-
-
63
Aclaude-opus-4-20250514
claude-opus-4-20250514Anthropic
23.5
1.4K
200K
¥108 / ¥540Input/Output
64
ST
step-1o-turbo-202506
step-1o-turbo-202506Stepfun
22.2
1.3K
-
-
65
MA
mistral-medium-2508
mistral-medium-2508Mistral
21.0
13.5K
262K
¥2.88 / ¥14.4Input/Output
66
Aclaude-sonnet-4-20250514
claude-sonnet-4-20250514Anthropic
19.8
1.1K
200K
¥21.6 / ¥108Input/Output
67
Zglm-4.6v
glm-4.6vZai
18.5
1.4K
128K
¥2.16 / ¥6.48Input/Output
68
STstep-3
step-3Stepfun
17.3
1.2K
65.5K
¥1.8 / ¥4.68Input/Output
69
TEhunyuan-large-vision
hunyuan-large-visionTencent
16.0
815
-
-
70
Ggemma-3-27b-it
gemma-3-27b-itGoogle
14.8
6.4K
128K
¥2.15 / ¥2.15Input/Output
71
Aclaude-3-7-sonnet-20250219
claude-3-7-sonnet-20250219Anthropic
13.6
915
200K
¥21.6 / ¥108Input/Output
72
Ogpt-5-nano-high
gpt-5-nano-highOpenai
12.3
1.7K
400K
¥0.36 / ¥2.88Input/Output
73
MAmistral-medium-2505
mistral-medium-2505Mistral
11.1
4.7K
262K
¥2.88 / ¥14.4Input/Output
74
Zglm-4.5v
glm-4.5vZai
9.9
1.2K
64K
¥4.32 / ¥13Input/Output
75
Ggemini-2.0-flash-001
gemini-2.0-flash-001Google
8.6
3.8K
1.05M
¥1.08 / ¥4.32Input/Output
76
Mllama-4-maverick-17b-128e-instruct
llama-4-maverick-17b-128e-instructMeta
7.4
3.2K
1M
¥1.8 / ¥6.26Input/Output
77
Aclaude-3-5-sonnet-20241022
claude-3-5-sonnet-20241022Anthropic
6.2
967
200K
¥21.6 / ¥108Input/Output
78
MAmistral-small-2506
mistral-small-2506Mistral
4.9
4.2K
262K
¥2.88 / ¥14.4Input/Output
79
MAmistral-small-3.1-24b-instruct-2503
mistral-small-3.1-24b-instruct-2503Mistral
3.7
7.4K
262K
¥2.88 / ¥14.4Input/Output
80
Mllama-4-scout-17b-16e-instruct
llama-4-scout-17b-16e-instructMeta
2.5
2.9K
128K
¥1.44 / ¥5.62Input/Output
81
Aclaude-3-5-haiku-20241022
claude-3-5-haiku-20241022Anthropic
1.2
934
200K
¥5.76 / ¥28.8Input/Output
82
AI
molmo-2-8b
molmo-2-8bAllenai
0.0
791
-
-
claude-opus-4-7 why it ranks first
claude-opus-4-7 ranks first with a percent score of 100.0 and 4.8K 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
OCR 识别排行榜看什么指标？
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。
OCR 识别模型应该怎么选？
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。