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 EnglishData Oct 19, 2026
Claude Opus 4.6 (High)Anthropic · 1324

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 · English

167 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 · English License
Score Rank Votes
1 MPhi-3 Vision 128K InstructMicrosoftOpen 837±23 167 880 MIT
2 MPhi-3.5 Vision InstructMicrosoftOpen 875±19 166 1,608 MIT
3 Claude 3 Haiku (20240307)Anthropic 949±15 164 8,449 Proprietary
4 OMiniCPM-V 2.6OpenBMBOpen 949±19 165 1,256 Apache 2.0
5 Nnvila-internal-15b-v1NVIDIAOpen 959±26 163 619 -
6 LLLaVA-OneVision Qwen2 72B OVOpen 962±22 162 841 Apache 2.0
7 CogVLM2 Llama3 Chat 19BZ.aiOpen 969±20 161 1,162 CogVLM2
8 LLLaVA 1.6 34BOpen 970±16 160 2,813 Apache 2.0
9 Claude 3 Sonnet (20240229)Anthropic 980±14 159 7,799 Proprietary
10 OInternVL2 4BOpenGVLabOpen 986±16 158 2,189 MIT
11 AAmazon Nova Pro 1.0Amazon 987±19 157 1,264 Proprietary
12 Llama 3.2 11B Vision InstructMetaOpen 997±15 156 2,575 Llama 3.2
13 AAmazon Nova Lite 1.0Amazon 999±20 154 1,048 Proprietary
14 Aya Vision 32BCohereOpen 999±30 155 447 CC-BY-NC-4.0
15 Qwen2-VL-7B-InstructAlibabaOpen 1006±14 153 3,028 Apache 2.0
16 0Yi Vision01.AI 1007±23 152 785 Proprietary
17 Llama 3.2 90B Vision InstructMetaOpen 1013±12 150 4,630 Llama 3.2
18 Molmo 7B-D (0924)Ai2Open 1013±18 151 1,530 Apache 2.0
19 Claude 3 Opus (20240229)Anthropic 1018±13 149 9,732 Proprietary
20 Hunyuan Standard Vision (2024-12-31)Tencent 1020±28 148 450 Proprietary
21 Gemini 1.5 Flash (001)Google 1027±14 146 8,419 Proprietary
22 Pixtral 12B (2409)Mistral AIOpen 1027±12 147 4,090 Apache 2.0
23 OInternVL2 26BOpenGVLabOpen 1037±15 145 3,250 MIT
24 Molmo 72B (0924)Ai2Open 1045±17 144 1,627 Apache 2.0
25 Qwen2-VL-72BAlibabaOpen 1054±13 143 3,230 Qwen
26 Gemini 1.5 Flash-8B (001)Google 1070±13 142 3,348 Proprietary
27 SStep-1V 32KStepFun 1071±22 141 839 Proprietary
28 GPT-4o mini (2024-07-18)OpenAI 1072±11 138 10,721 Proprietary
29 Qwen-VL-Max (1119)Alibaba 1072±22 139 755 Proprietary
30 GPT-4o (2024-08-06)OpenAI 1072±16 140 1,744 Proprietary
31 Claude 3.5 Haiku (20241022)Anthropic 1088±23 137 652 Proprietary
32 Gemini 1.5 Pro (001)Google 1089±14 136 10,587 Proprietary
33 Molmo 2 8BAi2Open 1092±29 135 461 Apache 2.0
34 GPT-4 Turbo (2024-04-09)OpenAI 1097±14 134 8,417 Proprietary
35 Pixtral Large (2411)Mistral AIOpen 1100±12 133 2,921 MRL
36 GPT-4.1 nano (2025-04-14)OpenAI 1101±27 132 461 Proprietary
37 Qwen2.5-VL-72B-InstructAlibabaOpen 1115±14 131 2,169 Qwen
38 Claude 3.5 Sonnet (20240620)Anthropic 1119±13 130 13,270 Proprietary
39 Gemini 2.0 Flash-Lite Preview (02-05)Google 1121±14 129 2,252 Proprietary
40 Claude 3.5 Sonnet (20241022)Anthropic 1124±10 128 5,352 Proprietary
41 Claude 3.7 Sonnet (20250219)Anthropic 1141±13 127 2,156 Proprietary
42 Llama 4 Scout 17B 16E InstructMetaOpen 1143±12 126 3,098 Llama
43 GPT-4o (2024-05-13)OpenAI 1147±11 124 14,077 Proprietary
44 SStep-1o Vision 32K (High Res)StepFun 1147±16 125 1,599 Proprietary
45 Llama 4 Maverick 17B 128E InstructMetaOpen 1150±12 123 3,387 Llama 4
46 Mistral Small 3.1 24B Instruct (2503)Mistral AIOpen 1157±10 121 14,751 Apache 2.0
47 Gemini 1.5 Flash (002)Google 1157±13 122 3,888 Proprietary
48 Qwen2.5-VL-32B-InstructAlibabaOpen 1158±19 119 870 Apache 2.0
49 Claude Sonnet 4 (20250514)Anthropic 1158±21 120 761 Proprietary
50 Gemini 1.5 Pro (002)Google 1159±12 118 4,829 Proprietary
51 o1 (2024-12-17)OpenAI 1162±14 116 1,943 Proprietary
52 Mistral Small 3.2 (2506)Mistral AIOpen 1162±11 117 5,376 Apache 2.0
53 Claude 3.7 Sonnet (Thinking 32K, 20250219)Anthropic 1164±23 115 644 Proprietary
54 GPT-5 nano (High)OpenAI 1170±15 114 1,783 Proprietary
55 Mistral Medium 3 (2505)Mistral AI 1172±11 113 5,311 Proprietary
56 Gemini 2.0 Flash (001)Google 1174±9 112 4,961 Proprietary
57 Mistral Medium 3.1 (2508)Mistral AI 1179±8 110 19,018 Proprietary
58 GLM-4.6VZ.aiOpen 1179±20 111 914 MIT
59 Claude Opus 4 (Thinking 16K, 20250514)Anthropic 1182±25 109 549 Proprietary
60 GLM-4.5VZ.aiOpen 1183±16 108 1,596 MIT
61 Claude Sonnet 4 (Thinking 32K, 20250514)Anthropic 1186±25 106 526 Proprietary
62 Claude Opus 4 (20250514)Anthropic 1186±19 107 954 Proprietary
63 Gemma 3 27B ITGoogleOpen 1188±10 105 7,938 Gemma
64 Hunyuan Vision 1.5 ThinkingTencent 1190±18 103 1,028 Proprietary
65 Gemini 2.5 Flash-Lite Preview (Thinking, 06-17)Google 1190±10 104 16,579 Proprietary
66 GPT-4.5 Preview (2025-02-27)OpenAI 1191±16 101 1,392 Proprietary
67 GPT-4.1 mini (2025-04-14)OpenAI 1191±10 102 19,159 Proprietary
68 Hunyuan Large VisionTencent 1198±23 100 609 Proprietary
69 SStep-3StepFunOpen 1199±15 99 1,616 Apache 2.0
70 o4-mini (2025-04-16)OpenAI 1200±9 98 19,813 Proprietary
71 Gemini 2.5 Flash-Lite Preview (Non-Thinking, 09-2025)Google 1201±14 97 1,917 Proprietary
72 GPT-5.4 nano (High)OpenAI 1202±8 96 10,341 Proprietary
73 GPT-6 Luna (Max)OpenAI 1204±18 95 1,264 Proprietary
74 GPT-4.1 (2025-04-14)OpenAI 1209±9 94 19,011 Proprietary
75 SStep-1o Turbo (202506)StepFun 1212±20 93 850 Proprietary
76 Qwen3-VL-235B-A22B-ThinkingAlibabaOpen 1215±18 92 972 Apache 2.0
77 GPT-5 mini (High)OpenAI 1216±10 91 13,428 Proprietary
78 GPT-5 (High)OpenAI 1218±9 90 15,898 Proprietary
79 Grok 4.1 Fast (Reasoning)xAI 1219±10 89 5,239 Proprietary
80 o3 (2025-04-16)OpenAI 1221±9 88 21,609 Proprietary
81 Grok 4 (0709)xAI 1222±10 87 13,981 Proprietary
82 Qwen-VL-Max (2025-08-13)Alibaba 1227±16 86 1,477 Proprietary
83 GPT-6 Sol (Max)OpenAI 1229±19 84 1,058 Proprietary
84 Grok 4.7 (xHigh)xAI 1229±18 85 1,292 Proprietary
85 GPT-5.1OpenAI 1230±11 82 3,727 Proprietary
86 GPT-5.2OpenAI 1230±9 83 7,703 Proprietary
87 GPT-5 ChatOpenAI 1231±10 81 18,691 Proprietary
88 Gemini 3.1 Flash-Lite PreviewGoogle 1232±8 80 11,005 Proprietary
89 Grok 4.3xAI 1235±8 78 11,160 Proprietary
90 Mistral Large 3Mistral AIOpen 1235±12 79 3,184 Apache 2.0
91 Gemini 2.5 FlashGoogle 1239±6 76 25,447 Proprietary
92 NNVIDIA-Nemotron-3.5-Super-VL-120B-A12B (NVFP4)NVIDIAOpen 1239±21 77 864 OpenMDW-1.1
93 Mistral Medium 3.5Mistral AIOpen 1240±14 75 2,314 Modified MIT
94 MiMo-V2-OmniXiaomi 1242±11 74 4,294 Proprietary
95 GPT-5.2 (High)OpenAI 1243±9 72 7,347 Proprietary
96 TInkling-SmallThinking MachinesOpen 1243±12 73 2,816 Apache 2.0
97 ChatGPT-4o Latest (20250326)OpenAI 1244±7 70 10,179 Proprietary
98 GPT-5.4 mini (High)OpenAI 1244±8 71 10,694 Proprietary
99 Qwen3.5-27BAlibabaOpen 1246±8 69 11,800 Apache 2.0
100 Qwen3.5-122B-A10BAlibabaOpen 1249±10 68 5,296 Apache 2.0
101 ERNIE-5.0-Preview (1220)Baidu 1250±18 67 1,180 Proprietary
102 GPT-5.1 (High)OpenAI 1252±11 66 3,532 Proprietary
103 GPT-5.5 InstantOpenAI 1254±12 65 3,328 Proprietary
104 Gemma 4 26B A4BGoogleOpen 1256±9 63 9,635 Apache 2.0
105 Qwen3-VL-235B-A22B-InstructAlibabaOpen 1256±10 64 4,605 Apache 2.0
106 Gemini 3 Flash (Minimal)Google 1258±7 61 14,728 Proprietary
107 Kimi K2.5 (Instant)Moonshot AIOpen 1258±16 62 1,419 Modified MIT
108 MiMo-V2.5XiaomiOpen 1259±8 60 11,479 MIT
109 Gemini 2.5 ProGoogle 1261±6 59 40,672 Proprietary
110 Grok 4.20 Multi-Agent Beta (0309)xAI 1262±8 58 9,842 Proprietary
111 MiniMax-M3MiniMaxOpen 1263±9 56 7,372 MiniMax Community License
112 GPT-5.6 Luna (xHigh)OpenAI 1263±11 57 3,794 Proprietary
113 Gemini 2.5 Flash Preview (09-2025)Google 1264±14 55 1,813 Proprietary
114 Grok 4.20 Beta (Reasoning, 0309)xAI 1266±8 54 10,741 Proprietary
115 Qwen3.8-27BAlibabaOpen 1267±13 53 2,445 Apache 2.0
116 Qwen3.5-397B-A17BAlibabaOpen 1268±7 51 12,659 Apache 2.0
117 MiMo-V2.6-ProXiaomiOpen 1268±24 52 612 MIT
118 MiMo-V2.6-FlashXiaomiOpen 1269±18 49 1,134 MIT
119 Gemini 3.5 Flash-LiteGoogle 1269±12 50 3,159 Proprietary
120 Grok 4.6 (High)xAI 1270±13 48 2,512 Proprietary
121 Gemma 4 31BGoogleOpen 1271±8 47 15,807 Apache 2.0
122 GPT-5.6 Terra (xHigh)OpenAI 1272±11 46 4,052 Proprietary
123 GLM-5V-TurboZ.ai 1274±7 43 17,081 Proprietary
124 Kimi K2.5 (Thinking)Moonshot AIOpen 1274±8 44 10,972 Modified MIT
125 GPT-6.1 Sol (Max)OpenAI 1274±22 45 774 Proprietary
126 GPT-5.2 Chat (20260210)OpenAI 1275±9 41 6,385 Proprietary
127 Claude Sonnet 5 (High)Anthropic 1275±10 42 5,335 Proprietary
128 Gemini 3 FlashGoogle 1276±7 40 15,117 Proprietary
129 Qwen3.7-PlusAlibaba 1280±10 37 5,067 Proprietary
130 GPT-5.6 Sol (xHigh)OpenAI 1280±11 38 3,912 Proprietary
131 GPT-6 Astra (Max)OpenAI 1280±17 39 1,343 Proprietary
132 Mistral Large 4Mistral AI 1283±31 36 355 Proprietary
133 Grok 4.5xAI 1285±10 35 4,285 Proprietary
134 Gemini 3.1 Pro PreviewGoogle 1287±7 33 18,497 Proprietary
135 Kimi K2.6Moonshot AIOpen 1287±9 34 6,756 Modified MIT
136 DeepSeek-V4.1-Flash-20260910 (Max)DeepSeekOpen 1288±17 32 1,301 MIT
137 Claude Sonnet 5.5 (xHigh)Anthropic 1289±21 31 859 Proprietary
138 SStep 5 PreviewStepFun 1290±26 29 513 Proprietary
139 GPT-5.4OpenAI 1290±9 30 9,595 Proprietary
140 GPT-5.5 (High)OpenAI 1291±8 26 10,772 Proprietary
141 Muse Spark 1.2 (xHigh)Meta 1291±23 27 695 Proprietary
142 Claude Opus 4.8 (High)Anthropic 1291±9 28 7,224 Proprietary
143 GPT-5.5OpenAI 1292±8 22 11,422 Proprietary
144 Muse Spark 1.1Meta 1292±10 23 4,266 Proprietary
145 Gemini 3 ProGoogle 1292±10 24 4,724 Proprietary
146 Muse Spark 1.3 (xHigh)Meta 1292±16 25 1,654 Proprietary
147 Claude Opus 4.8Anthropic 1293±9 21 7,311 Proprietary
148 Dola Seed 2.0 ProByteDance 1294±11 20 4,042 Proprietary
149 GLM-5.3-FlashZ.aiOpen 1295±10 18 5,043 MIT
150 Gemini 3.6 Flash (High)Google 1295±12 19 3,098 Proprietary
151 Claude Sonnet 4.6Anthropic 1298±8 16 11,555 Proprietary
152 Muse Spark 1.3 (Max)Meta 1298±15 17 1,877 Proprietary
153 Gemini 3.5 Flash (Medium)Google 1299±10 15 5,035 Proprietary
154 GPT-5.4 (High)OpenAI 1301±8 14 12,132 Proprietary
155 Muse SparkMeta 1304±13 13 2,428 Proprietary
156 Gemini 3.5 Flash (High)Google 1305±10 12 5,040 Proprietary
157 Gemini 3.7 Flash (High)Google 1309±15 11 1,848 Proprietary
158 Qwen3.8-MaxAlibaba 1311±11 10 3,284 Proprietary
159 Claude Fable 5.1 (Max)Anthropic 1313±15 9 1,825 Proprietary
160 Claude Opus 5 (High)Anthropic 1315±9 8 6,237 Proprietary
161 Claude Opus 5.5 (High)Anthropic 1317±20 7 992 Proprietary
162 Gemini 3.8 Flash (High)Google 1318±16 6 1,556 Proprietary
163 Claude Opus 4.6Anthropic 1319±8 5 11,334 Proprietary
164 Claude Opus 4.7 (High)Anthropic 1320±9 4 9,705 Proprietary
165 Claude Opus 4.7Anthropic 1322±9 3 9,834 Proprietary
166 Claude Fable 5 (High)Anthropic 1323±10 2 5,874 Proprietary
167 Claude Opus 4.6 (High)Anthropic 1324±9 1 9,235 Proprietary

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

This week

Collecting data: 1/7 days recorded; weekly changes appear after 7 days

Synced automatically once a day; ranks are compared only between real snapshots, never backfilled.

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.