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 HumorData Oct 19, 2026
Muse SparkMeta · 1344

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 · Humor

93 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 · Humor License
Score Rank Votes
1 Llama 4 Maverick 17B 128E InstructMetaOpen 1100±40 93 216 Llama 4
2 Llama 4 Scout 17B 16E InstructMetaOpen 1102±45 92 191 Llama
3 GPT-5 nano (High)OpenAI 1110±47 91 174 Proprietary
4 Mistral Small 3.2 (2506)Mistral AIOpen 1140±30 90 481 Apache 2.0
5 Mistral Small 3.1 24B Instruct (2503)Mistral AIOpen 1142±23 89 933 Apache 2.0
6 Mistral Medium 3 (2505)Mistral AI 1145±30 88 440 Proprietary
7 GPT-5.4 nano (High)OpenAI 1161±21 87 923 Proprietary
8 Gemini 2.0 Flash (001)Google 1165±41 86 227 Proprietary
9 Gemma 3 27B ITGoogleOpen 1166±24 85 762 Gemma
10 Mistral Medium 3.1 (2508)Mistral AI 1168±19 84 1,502 Proprietary
11 Gemini 2.5 Flash-Lite Preview (Non-Thinking, 09-2025)Google 1170±37 83 212 Proprietary
12 GPT-4.1 mini (2025-04-14)OpenAI 1182±21 82 1,403 Proprietary
13 MiMo-V2-OmniXiaomi 1184±32 81 370 Proprietary
14 Gemini 2.5 Flash-Lite Preview (Thinking, 06-17)Google 1187±21 80 1,387 Proprietary
15 GPT-5.6 Luna (xHigh)OpenAI 1190±27 79 502 Proprietary
16 ERNIE-5.0-Preview (1220)Baidu 1194±43 78 178 Proprietary
17 GPT-5.2OpenAI 1196±24 76 659 Proprietary
18 GPT-4.1 (2025-04-14)OpenAI 1196±21 77 1,365 Proprietary
19 o4-mini (2025-04-16)OpenAI 1199±20 75 1,449 Proprietary
20 GPT-5 mini (High)OpenAI 1205±22 74 1,047 Proprietary
21 TInkling-SmallThinking MachinesOpen 1207±30 72 384 Apache 2.0
22 GPT-5 (High)OpenAI 1207±21 73 1,318 Proprietary
23 Qwen3-VL-235B-A22B-InstructAlibabaOpen 1209±26 70 508 Apache 2.0
24 Grok 4 (0709)xAI 1209±21 71 1,135 Proprietary
25 Mistral Medium 3.5Mistral AIOpen 1216±39 69 238 Modified MIT
26 Mistral Large 3Mistral AIOpen 1218±29 68 452 Apache 2.0
27 GPT-5 ChatOpenAI 1219±20 67 1,448 Proprietary
28 o3 (2025-04-16)OpenAI 1222±19 65 1,643 Proprietary
29 Grok 4.3xAI 1222±19 66 1,069 Proprietary
30 Gemini 2.5 FlashGoogle 1225±15 64 2,192 Proprietary
31 Kimi K2.5 (Instant)Moonshot AIOpen 1226±46 62 155 Modified MIT
32 MiMo-V2.5XiaomiOpen 1226±18 63 1,147 MIT
33 GPT-5.1 (High)OpenAI 1227±31 59 340 Proprietary
34 Qwen3.5-27BAlibabaOpen 1227±18 60 1,195 Apache 2.0
35 GPT-5.2 (High)OpenAI 1227±24 61 681 Proprietary
36 Qwen3.8-27BAlibabaOpen 1235±32 56 347 Apache 2.0
37 GPT-5.4 mini (High)OpenAI 1235±21 57 868 Proprietary
38 MiniMax-M3MiniMaxOpen 1235±22 58 773 MiniMax Community License
39 GPT-5.6 Terra (xHigh)OpenAI 1236±26 55 526 Proprietary
40 Grok 4.1 Fast (Reasoning)xAI 1239±28 53 482 Proprietary
41 GPT-5.2 Chat (20260210)OpenAI 1239±26 54 552 Proprietary
42 Qwen3.5-122B-A10BAlibabaOpen 1243±27 52 468 Apache 2.0
43 GPT-5.6 Sol (xHigh)OpenAI 1246±25 51 569 Proprietary
44 Claude Sonnet 4.6Anthropic 1248±19 49 1,060 Proprietary
45 ChatGPT-4o Latest (20250326)OpenAI 1248±22 50 795 Proprietary
46 Gemini 2.5 Flash Preview (09-2025)Google 1255±37 48 216 Proprietary
47 DeepSeek-V4.1-Flash-20260910 (Max)DeepSeekOpen 1260±45 47 166 MIT
48 Qwen3.7-PlusAlibaba 1261±25 44 556 Proprietary
49 GPT-5.5 InstantOpenAI 1261±40 45 241 Proprietary
50 Claude Sonnet 5 (High)Anthropic 1261±24 46 628 Proprietary
51 Gemini 2.5 ProGoogle 1262±15 41 3,133 Proprietary
52 Qwen3.5-397B-A17BAlibabaOpen 1262±17 42 1,223 Apache 2.0
53 GPT-5.4OpenAI 1262±21 43 844 Proprietary
54 GPT-5.1OpenAI 1263±32 40 376 Proprietary
55 Gemini 3.1 Flash-Lite PreviewGoogle 1266±19 39 979 Proprietary
56 GLM-5V-TurboZ.ai 1268±15 38 1,765 Proprietary
57 Grok 4.6 (High)xAI 1270±33 37 320 Proprietary
58 Gemma 4 26B A4BGoogleOpen 1272±19 36 993 Apache 2.0
59 Dola Seed 2.0 ProByteDance 1275±30 35 393 Proprietary
60 Grok 4.5xAI 1276±25 34 573 Proprietary
61 Gemini 3 Flash (Minimal)Google 1277±16 32 1,389 Proprietary
62 GPT-5.5OpenAI 1277±18 33 1,130 Proprietary
63 GLM-5.3-FlashZ.aiOpen 1279±23 30 671 MIT
64 Grok 4.20 Beta (Reasoning, 0309)xAI 1279±19 31 1,003 Proprietary
65 Kimi K2.6Moonshot AIOpen 1280±25 28 590 Modified MIT
66 Gemma 4 31BGoogleOpen 1280±15 29 1,659 Apache 2.0
67 Claude Opus 4.8Anthropic 1281±22 26 783 Proprietary
68 Grok 4.20 Multi-Agent Beta (0309)xAI 1281±20 27 935 Proprietary
69 Claude Opus 4.8 (High)Anthropic 1282±22 25 756 Proprietary
70 GPT-5.4 (High)OpenAI 1284±18 23 1,143 Proprietary
71 Kimi K2.5 (Thinking)Moonshot AIOpen 1284±18 24 1,083 Modified MIT
72 Gemini 3.5 Flash-LiteGoogle 1287±30 22 396 Proprietary
73 Muse Spark 1.1Meta 1288±25 20 577 Proprietary
74 Muse Spark 1.3 (Max)Meta 1288±39 21 222 Proprietary
75 Gemini 3.6 Flash (High)Google 1289±29 19 427 Proprietary
76 Gemini 3 FlashGoogle 1291±16 18 1,442 Proprietary
77 Claude Opus 5 (High)Anthropic 1293±21 15 876 Proprietary
78 Muse Spark 1.3 (xHigh)Meta 1293±42 16 208 Proprietary
79 GPT-5.5 (High)OpenAI 1293±18 17 1,075 Proprietary
80 Gemini 3.7 Flash (High)Google 1299±41 14 215 Proprietary
81 Qwen3.8-MaxAlibaba 1304±29 13 435 Proprietary
82 Claude Opus 4.7Anthropic 1305±20 12 896 Proprietary
83 Claude Opus 4.7 (High)Anthropic 1309±21 11 889 Proprietary
84 Gemini 3.5 Flash (Medium)Google 1313±26 10 565 Proprietary
85 Claude Opus 4.6Anthropic 1314±19 8 1,026 Proprietary
86 Claude Opus 4.6 (High)Anthropic 1314±21 9 867 Proprietary
87 Claude Fable 5.1 (Max)Anthropic 1318±40 7 208 Proprietary
88 Gemini 3.5 Flash (High)Google 1321±25 5 577 Proprietary
89 Gemini 3.8 Flash (High)Google 1321±45 6 175 Proprietary
90 Gemini 3.1 Pro PreviewGoogle 1322±14 4 1,885 Proprietary
91 Gemini 3 ProGoogle 1339±25 3 611 Proprietary
92 Claude Fable 5 (High)Anthropic 1341±25 2 583 Proprietary
93 Muse SparkMeta 1344±40 1 237 Proprietary

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