Vision Model Leaderboard: Best Vision-Language Models

Ranked by LMArena Vision Arena scores: users upload an image, ask a question, and vote between two anonymous models’ answers.

Overall OCR Diagrams Homework Captioning Entity recognition Creative writing Humor English Chinese
#1 ChineseData Oct 19, 2026
Gemini 3.8 Flash (High)Google · 1416

In LMArena Vision Arena data published Oct 19, 2026, Anthropic’s Claude Fable 5 (High) ranks #1 of 167 models with 1325 (14,714 votes).

Full leaderboard · Chinese

128 models
All Open weights

Sorted by LMArena score by default; ± is half the 95% confidence interval. Category columns show the model’s score on that sub-board; – means no score.

# Model LMArena · Chinese License
Score Rank Votes
1 Gemini 3.8 Flash (High)Google 1416±39 1 224 Proprietary
2 Claude Opus 5 (High)Anthropic 1400±20 2 1,037 Proprietary
3 Gemini 3.7 Flash (High)Google 1394±34 3 302 Proprietary
4 Claude Fable 5.1 (Max)Anthropic 1391±35 4 284 Proprietary
5 Muse Spark 1.3 (xHigh)Meta 1384±36 5 257 Proprietary
6 Gemini 3.5 Flash (High)Google 1377±22 6 847 Proprietary
7 Gemini 3.5 Flash (Medium)Google 1375±22 7 818 Proprietary
8 Claude Fable 5 (High)Anthropic 1373±23 8 784 Proprietary
9 Claude Opus 4.7Anthropic 1371±18 9 1,504 Proprietary
10 Gemini 3 ProGoogle 1370±17 10 1,687 Proprietary
11 Claude Opus 4.6Anthropic 1370±17 11 1,681 Proprietary
12 Muse Spark 1.3 (Max)Meta 1368±34 12 290 Proprietary
13 Claude Opus 4.6 (High)Anthropic 1367±18 13 1,487 Proprietary
14 Gemini 3.1 Pro PreviewGoogle 1364±13 14 3,186 Proprietary
15 GLM-5.3-FlashZ.aiOpen 1363±23 15 699 MIT
16 GPT-6 Astra (Max)OpenAI 1361±37 16 259 Proprietary
17 GPT-5.5OpenAI 1360±17 17 1,777 Proprietary
18 Qwen3.8-MaxAlibaba 1358±26 18 541 Proprietary
19 Gemini 3.6 Flash (High)Google 1357±26 19 536 Proprietary
20 GPT-5.5 (High)OpenAI 1357±17 20 1,603 Proprietary
21 Claude Opus 4.7 (High)Anthropic 1355±18 21 1,388 Proprietary
22 GPT-5.4 (High)OpenAI 1354±16 22 1,846 Proprietary
23 Muse Spark 1.1Meta 1348±22 23 767 Proprietary
24 Muse SparkMeta 1347±28 24 493 Proprietary
25 Kimi K2.6Moonshot AIOpen 1346±20 25 1,027 Modified MIT
26 Claude Opus 4.8Anthropic 1346±20 26 1,131 Proprietary
27 Grok 4.5xAI 1345±23 27 728 Proprietary
28 Claude Opus 4.8 (High)Anthropic 1342±20 28 1,030 Proprietary
29 Dola Seed 2.0 ProByteDance 1337±22 29 869 Proprietary
30 GPT-5.6 Terra (xHigh)OpenAI 1337±24 30 682 Proprietary
31 GPT-5.4OpenAI 1336±19 31 1,356 Proprietary
32 Qwen3.7-PlusAlibaba 1332±23 32 740 Proprietary
33 Claude Sonnet 5 (High)Anthropic 1331±22 33 847 Proprietary
34 GLM-5V-TurboZ.ai 1329±14 34 2,644 Proprietary
35 Gemini 3 FlashGoogle 1329±14 35 2,749 Proprietary
36 Kimi K2.5 (Instant)Moonshot AIOpen 1328±34 36 273 Modified MIT
37 DeepSeek-V4.1-Flash-20260910 (Max)DeepSeekOpen 1327±42 37 199 MIT
38 Gemma 4 31BGoogleOpen 1326±14 38 2,734 Apache 2.0
39 MiMo-V2.5XiaomiOpen 1326±17 39 1,580 MIT
40 Qwen3.8-27BAlibabaOpen 1326±30 40 409 Apache 2.0
41 Kimi K2.5 (Thinking)Moonshot AIOpen 1324±16 41 1,767 Modified MIT
42 Gemini 3 Flash (Minimal)Google 1317±14 42 2,539 Proprietary
43 Qwen3.5-397B-A17BAlibabaOpen 1314±15 43 2,045 Apache 2.0
44 GPT-5.6 Sol (xHigh)OpenAI 1313±24 44 677 Proprietary
45 Gemma 4 26B A4BGoogleOpen 1312±17 45 1,609 Apache 2.0
46 MiniMax-M3MiniMaxOpen 1311±20 46 1,048 MiniMax Community License
47 Gemini 3.5 Flash-LiteGoogle 1311±27 47 524 Proprietary
48 Claude Sonnet 4.6Anthropic 1307±17 48 1,736 Proprietary
49 GPT-5.2 (High)OpenAI 1305±18 49 1,351 Proprietary
50 Grok 4.20 Beta (Reasoning, 0309)xAI 1304±16 50 1,833 Proprietary
51 Qwen3.5-122B-A10BAlibabaOpen 1301±20 51 1,011 Apache 2.0
52 Qwen3.5-27BAlibabaOpen 1301±15 52 1,997 Apache 2.0
53 Gemini 2.5 ProGoogle 1300±13 53 4,797 Proprietary
54 Grok 4.20 Multi-Agent Beta (0309)xAI 1300±17 54 1,722 Proprietary
55 GPT-5.6 Luna (xHigh)OpenAI 1294±25 55 631 Proprietary
56 TInkling-SmallThinking MachinesOpen 1292±28 56 461 Apache 2.0
57 Grok 4.6 (High)xAI 1289±31 57 380 Proprietary
58 Gemini 2.5 Flash Preview (09-2025)Google 1288±27 58 460 Proprietary
59 GPT-5.4 mini (High)OpenAI 1286±17 59 1,661 Proprietary
60 Gemini 2.5 FlashGoogle 1284±13 60 3,692 Proprietary
61 MiMo-V2-OmniXiaomi 1284±24 61 693 Proprietary
62 Qwen3-VL-235B-A22B-InstructAlibabaOpen 1283±19 62 998 Apache 2.0
63 GPT-5.5 InstantOpenAI 1278±29 63 494 Proprietary
64 Gemini 3.1 Flash-Lite PreviewGoogle 1278±16 64 1,989 Proprietary
65 GPT-5.2OpenAI 1270±17 65 1,437 Proprietary
66 GPT-5.1 (High)OpenAI 1270±23 66 671 Proprietary
67 GPT-5.1OpenAI 1270±21 67 813 Proprietary
68 ERNIE-5.0-Preview (1220)Baidu 1268±37 68 253 Proprietary
69 GPT-5 ChatOpenAI 1267±17 69 2,052 Proprietary
70 GPT-6 Luna (Max)OpenAI 1266±44 70 169 Proprietary
71 GPT-5.2 Chat (20260210)OpenAI 1264±20 71 1,148 Proprietary
72 Grok 4.3xAI 1257±17 72 1,650 Proprietary
73 ChatGPT-4o Latest (20250326)OpenAI 1255±17 73 1,419 Proprietary
74 Grok 4.7 (xHigh)xAI 1250±45 74 182 Proprietary
75 Grok 4 (0709)xAI 1249±18 75 1,675 Proprietary
76 Mistral Medium 3.5Mistral AIOpen 1244±38 76 287 Modified MIT
77 Mistral Large 3Mistral AIOpen 1240±26 77 569 Apache 2.0
78 o3 (2025-04-16)OpenAI 1237±16 78 2,258 Proprietary
79 o1 (2024-12-17)OpenAI 1235±58 79 99 Proprietary
80 GPT-5 (High)OpenAI 1234±17 80 2,135 Proprietary
81 Grok 4.1 Fast (Reasoning)xAI 1233±21 81 1,062 Proprietary
82 Gemini 2.5 Flash-Lite Preview (Non-Thinking, 09-2025)Google 1233±25 82 490 Proprietary
83 GPT-5.4 nano (High)OpenAI 1232±17 83 1,739 Proprietary
84 GPT-4.1 (2025-04-14)OpenAI 1226±17 84 1,978 Proprietary
85 Qwen3-VL-235B-A22B-ThinkingAlibabaOpen 1219±43 85 164 Apache 2.0
86 o4-mini (2025-04-16)OpenAI 1218±17 86 1,934 Proprietary
87 GLM-4.6VZ.aiOpen 1217±41 87 166 MIT
88 GPT-5 mini (High)OpenAI 1214±18 88 1,600 Proprietary
89 Gemini 2.5 Flash-Lite Preview (Thinking, 06-17)Google 1194±19 89 1,655 Proprietary
90 Mistral Medium 3.1 (2508)Mistral AI 1190±16 90 2,176 Proprietary
91 Gemini 1.5 Pro (002)Google 1187±44 91 395 Proprietary
92 GPT-4.1 mini (2025-04-14)OpenAI 1186±18 92 1,778 Proprietary
93 GPT-4.5 Preview (2025-02-27)OpenAI 1176±63 93 77 Proprietary
94 GPT-5 nano (High)OpenAI 1176±43 94 189 Proprietary
95 Claude 3.7 Sonnet (20250219)Anthropic 1174±47 95 163 Proprietary
96 Gemma 3 27B ITGoogleOpen 1164±21 96 1,020 Gemma
97 Mistral Small 3.2 (2506)Mistral AIOpen 1163±23 97 847 Apache 2.0
98 Mistral Medium 3 (2505)Mistral AI 1161±25 98 599 Proprietary
99 Mistral Small 3.1 24B Instruct (2503)Mistral AIOpen 1155±21 99 1,251 Apache 2.0
100 Gemini 2.0 Flash (001)Google 1145±30 100 414 Proprietary
101 Llama 4 Maverick 17B 128E InstructMetaOpen 1141±34 101 294 Llama 4
102 Gemini 1.5 Flash (002)Google 1137±44 102 370 Proprietary
103 GPT-4o (2024-05-13)OpenAI 1132±38 103 1,472 Proprietary
104 Claude 3.5 Sonnet (20241022)Anthropic 1119±38 104 514 Proprietary
105 Qwen2.5-VL-72B-InstructAlibabaOpen 1107±55 105 119 Qwen
106 Llama 4 Scout 17B 16E InstructMetaOpen 1106±36 106 278 Llama
107 Claude 3.5 Sonnet (20240620)Anthropic 1105±39 107 1,564 Proprietary
108 Pixtral Large (2411)Mistral AIOpen 1083±51 108 156 MRL
109 Qwen2-VL-72BAlibabaOpen 1072±45 109 291 Qwen
110 OInternVL2 26BOpenGVLabOpen 1065±48 110 287 MIT
111 GPT-4 Turbo (2024-04-09)OpenAI 1045±41 111 937 Proprietary
112 GPT-4o mini (2024-07-18)OpenAI 1033±38 112 954 Proprietary
113 Gemini 1.5 Pro (001)Google 1030±40 113 1,168 Proprietary
114 Gemini 2.0 Flash-Lite Preview (02-05)Google 1026±62 114 98 Proprietary
115 GPT-4o (2024-08-06)OpenAI 1009±52 115 186 Proprietary
116 Qwen2-VL-7B-InstructAlibabaOpen 999±46 116 298 Apache 2.0
117 Claude 3 Opus (20240229)Anthropic 994±40 117 1,038 Proprietary
118 Gemini 1.5 Flash-8B (001)Google 974±45 118 332 Proprietary
119 Gemini 1.5 Flash (001)Google 960±41 119 950 Proprietary
120 Llama 3.2 90B Vision InstructMetaOpen 952±43 120 393 Llama 3.2
121 Pixtral 12B (2409)Mistral AIOpen 951±45 121 335 Apache 2.0
122 OInternVL2 4BOpenGVLabOpen 950±55 122 177 MIT
123 Claude 3 Sonnet (20240229)Anthropic 942±41 123 915 Proprietary
124 Molmo 72B (0924)Ai2Open 936±55 124 182 Apache 2.0
125 Claude 3 Haiku (20240307)Anthropic 916±41 125 1,019 Proprietary
126 LLLaVA 1.6 34BOpen 915±47 126 415 Apache 2.0
127 Llama 3.2 11B Vision InstructMetaOpen 915±49 127 266 Llama 3.2
128 Molmo 7B-D (0924)Ai2Open 890±54 128 171 Apache 2.0

Scores: LMArena leaderboard-dataset (lmarena-ai on Hugging Face, CC-BY-4.0, vision config), Chinese published Oct 19, 2026. Please credit the source when citing.

Sources & methodology

  1. Scores are Bradley-Terry Arena ratings fitted from human votes between two anonymous answers to the same image prompt.
  2. Categories (OCR, diagrams, homework, captioning, Chinese/English…) are separate LMArena boards computed on prompts of that type, not a breakdown of the overall score; ranks can differ a lot.
  3. Fields without public data (price, context length, release date…) are not shown; blanks are –. Nothing is estimated.
  4. Synced daily. JSON API: /api/public/vision-leaderboard.php

FAQ

What is the best vision-language model right now?

In LMArena Vision Arena data published Oct 19, 2026, Anthropic’s Claude Fable 5 (High) ranks #1 of 167 models with 1325 (14,714 votes).

What is each vendor’s best vision model?

Top overall entry per vendor: Anthropic Claude Fable 5 (High) (1325, #1), Google Gemini 3.8 Flash (High) (1315, #7), Alibaba Qwen3.8-Max (1313, #10), Meta Muse Spark 1.3 (Max) (1308, #12), OpenAI GPT-5.4 (High) (1303, #18).

Which model is best at OCR?

On LMArena’s OCR board (123 models), Claude Fable 5 (High) is #1 with 1337.

What is the best open-weight vision model?

50 open-weight models are listed; the top one is GLM-5.3-Flash (1294, MIT).

How is this different from benchmarks like MMMU?

MMMU and similar benchmarks auto-score a fixed question set; this board comes from blind human votes on people’s own images and questions, reflecting real-use preference. They are not convertible.