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 OCRData Oct 19, 2026
Claude Fable 5 (High)Anthropic · 1337

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 · OCR

123 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 · OCR License
Score Rank Votes
1 Claude Fable 5 (High)Anthropic 1337±7 1 10,497 Proprietary
2 Gemini 3.8 Flash (High)Google 1333±11 2 3,137 Proprietary
3 Claude Opus 5 (High)Anthropic 1329±7 3 12,218 Proprietary
4 Claude Fable 5.1 (Max)Anthropic 1327±10 4 3,456 Proprietary
5 Claude Opus 4.7Anthropic 1327±7 5 16,736 Proprietary
6 Claude Opus 4.6Anthropic 1326±6 6 19,407 Proprietary
7 Claude Opus 4.6 (High)Anthropic 1326±7 7 15,908 Proprietary
8 Claude Opus 5.5 (High)Anthropic 1326±14 8 2,118 Proprietary
9 Qwen3.8-MaxAlibaba 1322±8 9 6,896 Proprietary
10 Gemini 3.7 Flash (High)Google 1322±10 10 3,653 Proprietary
11 Claude Opus 4.7 (High)Anthropic 1320±7 11 16,364 Proprietary
12 Muse Spark 1.3 (Max)Meta 1319±10 12 3,701 Proprietary
13 Gemini 3.5 Flash (Medium)Google 1317±7 13 9,712 Proprietary
14 Muse Spark 1.2 (xHigh)Meta 1313±17 14 1,343 Proprietary
15 Gemini 3.5 Flash (High)Google 1312±7 15 9,583 Proprietary
16 GPT-5.4 (High)OpenAI 1312±6 16 21,790 Proprietary
17 Gemini 3.6 Flash (High)Google 1312±8 17 6,075 Proprietary
18 Gemini 3 ProGoogle 1310±8 18 8,321 Proprietary
19 Muse Spark 1.3 (xHigh)Meta 1309±11 19 3,280 Proprietary
20 Muse SparkMeta 1307±10 20 4,223 Proprietary
21 GPT-5.5OpenAI 1306±6 21 20,286 Proprietary
22 GPT-5.5 (High)OpenAI 1304±6 22 18,851 Proprietary
23 Gemini 3.1 Pro PreviewGoogle 1304±5 23 32,719 Proprietary
24 Claude Opus 4.8 (High)Anthropic 1303±7 24 12,994 Proprietary
25 Muse Spark 1.1Meta 1301±7 25 8,699 Proprietary
26 GLM-5.3-FlashZ.aiOpen 1299±7 26 11,021 MIT
27 GPT-5.4OpenAI 1299±7 27 16,434 Proprietary
28 Claude Opus 4.8Anthropic 1298±7 28 13,248 Proprietary
29 Claude Sonnet 5.5 (xHigh)Anthropic 1297±14 29 1,780 Proprietary
30 Claude Sonnet 4.6Anthropic 1296±6 30 19,839 Proprietary
31 Grok 4.5xAI 1293±7 31 8,562 Proprietary
32 GPT-6 Astra (Max)OpenAI 1292±12 32 2,740 Proprietary
33 Kimi K2.6Moonshot AIOpen 1291±7 33 11,758 Modified MIT
34 Gemini 3 FlashGoogle 1291±5 34 26,690 Proprietary
35 GPT-5.6 Sol (xHigh)OpenAI 1290±8 35 7,710 Proprietary
36 GPT-6.1 Sol (Max)OpenAI 1289±15 36 1,654 Proprietary
37 Qwen3.7-PlusAlibaba 1288±7 37 9,923 Proprietary
38 Claude Sonnet 5 (High)Anthropic 1288±7 38 9,827 Proprietary
39 Mistral Large 4Mistral AI 1284±23 39 671 Proprietary
40 Dola Seed 2.0 ProByteDance 1284±8 40 7,715 Proprietary
41 DeepSeek-V4.1-Flash-20260910 (Max)DeepSeekOpen 1282±12 41 2,786 MIT
42 SStep 5 PreviewStepFun 1282±18 42 1,102 Proprietary
43 Gemma 4 31BGoogleOpen 1282±6 43 28,054 Apache 2.0
44 Qwen3.8-27BAlibabaOpen 1281±9 44 5,124 Apache 2.0
45 Gemini 3.5 Flash-LiteGoogle 1280±8 45 6,262 Proprietary
46 GPT-5.6 Terra (xHigh)OpenAI 1278±8 46 8,104 Proprietary
47 Kimi K2.5 (Thinking)Moonshot AIOpen 1276±6 47 19,489 Modified MIT
48 GPT-5.2 Chat (20260210)OpenAI 1275±7 48 11,343 Proprietary
49 Gemini 3 Flash (Minimal)Google 1272±5 49 25,665 Proprietary
50 Grok 4.6 (High)xAI 1272±9 50 4,881 Proprietary
51 Qwen3.5-397B-A17BAlibabaOpen 1271±5 51 22,696 Apache 2.0
52 Gemini 2.5 ProGoogle 1270±5 52 32,875 Proprietary
53 GLM-5V-TurboZ.ai 1270±5 53 30,138 Proprietary
54 MiMo-V2.6-FlashXiaomiOpen 1270±13 54 2,504 MIT
55 MiMo-V2.6-ProXiaomiOpen 1270±17 55 1,195 MIT
56 Gemma 4 26B A4BGoogleOpen 1268±7 56 16,862 Apache 2.0
57 Grok 4.20 Beta (Reasoning, 0309)xAI 1266±6 57 19,061 Proprietary
58 GPT-5.6 Luna (xHigh)OpenAI 1264±8 58 7,472 Proprietary
59 Grok 4.20 Multi-Agent Beta (0309)xAI 1261±6 59 17,215 Proprietary
60 Kimi K2.5 (Instant)Moonshot AIOpen 1259±12 60 2,691 Modified MIT
61 MiniMax-M3MiniMaxOpen 1259±6 61 13,947 MiniMax Community License
62 GPT-6 Sol (Max)OpenAI 1258±13 62 2,285 Proprietary
63 GPT-5.2 (High)OpenAI 1257±6 63 12,905 Proprietary
64 MiMo-V2.5XiaomiOpen 1256±6 64 20,901 MIT
65 GPT-5.5 InstantOpenAI 1256±9 65 5,653 Proprietary
66 GPT-5.1 (High)OpenAI 1255±9 66 6,005 Proprietary
67 Gemini 2.5 Flash Preview (09-2025)Google 1255±12 67 2,752 Proprietary
68 GPT-5.4 mini (High)OpenAI 1253±6 68 18,272 Proprietary
69 Qwen3-VL-235B-A22B-InstructAlibabaOpen 1253±8 69 7,752 Apache 2.0
70 Qwen3.5-122B-A10BAlibabaOpen 1251±7 70 9,185 Apache 2.0
71 ERNIE-5.0-Preview (1220)Baidu 1250±14 71 1,921 Proprietary
72 ChatGPT-4o Latest (20250326)OpenAI 1249±7 72 11,175 Proprietary
73 Gemini 3.1 Flash-Lite PreviewGoogle 1246±6 73 19,353 Proprietary
74 GPT-5 ChatOpenAI 1246±9 74 10,674 Proprietary
75 Qwen3.5-27BAlibabaOpen 1246±5 75 21,340 Apache 2.0
76 TInkling-SmallThinking MachinesOpen 1243±9 76 5,655 Apache 2.0
77 Grok 4.7 (xHigh)xAI 1241±12 77 2,508 Proprietary
78 GPT-5.1OpenAI 1241±8 78 6,725 Proprietary
79 MiMo-V2-OmniXiaomi 1238±8 79 7,549 Proprietary
80 Gemini 2.5 FlashGoogle 1237±5 80 27,156 Proprietary
81 Qwen-VL-Max (2025-08-13)Alibaba 1236±18 81 1,120 Proprietary
82 Grok 4.3xAI 1235±6 82 19,659 Proprietary
83 GPT-5.2OpenAI 1233±6 83 13,789 Proprietary
84 Mistral Medium 3.5Mistral AIOpen 1233±10 84 3,905 Modified MIT
85 Mistral Large 3Mistral AIOpen 1228±8 85 6,194 Apache 2.0
86 GPT-6 Luna (Max)OpenAI 1225±12 86 2,489 Proprietary
87 o3 (2025-04-16)OpenAI 1218±8 87 14,380 Proprietary
88 GPT-4.1 (2025-04-14)OpenAI 1218±9 88 11,355 Proprietary
89 NNVIDIA-Nemotron-3.5-Super-VL-120B-A12B (NVFP4)NVIDIAOpen 1216±15 89 1,672 OpenMDW-1.1
90 GPT-5 (High)OpenAI 1215±9 90 10,702 Proprietary
91 Qwen3-VL-235B-A22B-ThinkingAlibabaOpen 1212±16 91 1,322 Apache 2.0
92 GPT-5.4 nano (High)OpenAI 1204±6 92 18,222 Proprietary
93 Gemini 2.5 Flash-Lite Preview (Non-Thinking, 09-2025)Google 1203±12 93 2,787 Proprietary
94 GPT-5 mini (High)OpenAI 1199±10 94 7,999 Proprietary
95 o4-mini (2025-04-16)OpenAI 1197±9 95 11,774 Proprietary
96 Claude Sonnet 4 (Thinking 32K, 20250514)Anthropic 1197±21 96 725 Proprietary
97 Claude Opus 4 (Thinking 16K, 20250514)Anthropic 1197±20 97 835 Proprietary
98 Grok 4 (0709)xAI 1195±9 98 9,607 Proprietary
99 Grok 4.1 Fast (Reasoning)xAI 1195±7 99 9,260 Proprietary
100 Gemini 2.5 Flash-Lite Preview (Thinking, 06-17)Google 1188±9 100 10,500 Proprietary
101 GPT-4.1 mini (2025-04-14)OpenAI 1188±9 101 10,464 Proprietary
102 Claude 3.7 Sonnet (Thinking 32K, 20250219)Anthropic 1181±20 102 863 Proprietary
103 Claude Opus 4 (20250514)Anthropic 1179±16 103 1,331 Proprietary
104 SStep-1o Turbo (202506)StepFun 1177±17 104 1,244 Proprietary
105 Hunyuan Vision 1.5 ThinkingTencent 1176±16 105 1,364 Proprietary
106 Mistral Medium 3.1 (2508)Mistral AI 1175±7 106 13,469 Proprietary
107 Claude Sonnet 4 (20250514)Anthropic 1173±18 107 1,108 Proprietary
108 GLM-4.6VZ.aiOpen 1171±16 108 1,484 MIT
109 SStep-3StepFunOpen 1168±17 109 1,202 Apache 2.0
110 Hunyuan Large VisionTencent 1167±20 110 792 Proprietary
111 Claude 3.7 Sonnet (20250219)Anthropic 1157±19 111 879 Proprietary
112 Gemma 3 27B ITGoogleOpen 1156±10 112 6,196 Gemma
113 Mistral Medium 3 (2505)Mistral AI 1154±10 113 4,562 Proprietary
114 GPT-5 nano (High)OpenAI 1152±15 114 1,673 Proprietary
115 GLM-4.5VZ.aiOpen 1151±17 115 1,195 MIT
116 Gemini 2.0 Flash (001)Google 1148±12 116 3,690 Proprietary
117 Llama 4 Maverick 17B 128E InstructMetaOpen 1143±12 117 3,083 Llama 4
118 Claude 3.5 Sonnet (20241022)Anthropic 1139±19 118 943 Proprietary
119 Mistral Small 3.2 (2506)Mistral AIOpen 1134±11 119 4,083 Apache 2.0
120 Mistral Small 3.1 24B Instruct (2503)Mistral AIOpen 1132±11 120 7,209 Apache 2.0
121 Llama 4 Scout 17B 16E InstructMetaOpen 1117±12 121 2,788 Llama
122 Claude 3.5 Haiku (20241022)Anthropic 1091±20 122 905 Proprietary
123 Molmo 2 8BAi2Open 1058±24 123 796 Apache 2.0

Scores: LMArena leaderboard-dataset (lmarena-ai on Hugging Face, CC-BY-4.0, vision config), OCR 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.