Valumigo

LMArena · Image understanding (vision)

We regularly fetch and display public benchmark data: LMArena user-voted rankings and Epoch AI test scores. Each leaderboard shows its publication and retrieval dates. These scores are not produced by this site.

LMArena scores come from people comparing two models' answers side by side and voting for the better one, using an Elo-based system. If score differences fall within the confidence intervals, treat the models as roughly comparable.

How to read this leaderboard

What it measures: An image understanding ranking based on votes comparing two models' answers to questions that include images.

How to read the score: A relative score reflecting which answers users preferred. It is not an objective accuracy rate for interpreting photos, charts or screenshots.

Caveats: Check accuracy separately for the conditions you need, such as reading small text in documents or recognizing Korean (Hangul) characters.

Insights from this ranking

  • The 95% confidence intervals for 1st-place claude-fable-5-high and 2nd-place claude-opus-5-high overlap. This aggregation alone does not clearly establish their order.
  • 11 other models have confidence intervals that overlap with the 1st-place model's. Interpret small ranking differences cautiously alongside vote counts.
  • Anthropic has the most models among the top 10, with 7.
  • The 1st-place model received 13,709 votes, and this leaderboard includes 50 models.

Image understanding (vision) LMArena

Published 2026-10-02 · Retrieved 2026-10-04

1280130013201340
Anthropicclaude-fable-5-high
1325
Anthropicclaude-opus-5-high
1321
Anthropicclaude-fable-5.1-max
1320
Anthropicclaude-opus-4-7
1316
Googlegemini-3.8-flash-high
1316
Anthropicclaude-opus-4-6-high
1316
Googlegemini-3.7-flash-high
1315
Alibabaqwen3.8-max
1314
Anthropicclaude-opus-4-6
1313
Anthropicclaude-opus-4-7-high
1313
Metamuse-spark-1.3-max
1310
Googlegemini-3.5-flash-high
1310
Googlegemini-3.5-flash-medium
1308
Metamuse-spark
1305
Googlegemini-3-pro
1305
Metamuse-spark-1.2 (xHigh)
1305
Ogpt-5.4-high
1302
Zglm-5.3-flash
1301
Ogpt-5.5
1297
Googlegemini-3.6-flash-high
1297

Dots show scores; horizontal lines show 95% confidence intervals. Overlapping intervals suggest similar performance.

View table (top 50)
RankModelDeveloperScore95% confidence intervalVotes
1Anthropicclaude-fable-5-highAnthropic13251318–133213,709
2Anthropicclaude-opus-5-highAnthropic13211314–132815,900
3Anthropicclaude-fable-5.1-maxAnthropic13201310–13304,312
4Anthropicclaude-opus-4-7Anthropic13161310–132323,169
5Googlegemini-3.8-flash-highGoogle13161304–13273,086
6Anthropicclaude-opus-4-6-highAnthropic13161309–132222,328
7Googlegemini-3.7-flash-highGoogle13151305–13253,814
8Alibabaqwen3.8-maxAlibaba13141307–132210,040
9Anthropicclaude-opus-4-6Anthropic13131307–131926,869
10Anthropicclaude-opus-4-7-highAnthropic13131306–131922,755
11Metamuse-spark-1.3-maxMeta13101300–13213,843
12Googlegemini-3.5-flash-highGoogle13101303–131712,336
13Googlegemini-3.5-flash-mediumGoogle13081301–131512,467
14Metamuse-sparkMeta13051296–13145,868
15Googlegemini-3-proGoogle13051298–131213,620
16Metamuse-spark-1.2 (xHigh)Meta13051290–13191,951
17Ogpt-5.4-highOpenAI13021296–130829,177
18Zglm-5.3-flashZ.ai13011292–13105,479
19Ogpt-5.5OpenAI12971291–130327,223
20Googlegemini-3.6-flash-highGoogle12971288–13057,392
21Googlegemini-3.1-pro-previewGoogle12961291–130145,558
22Anthropicclaude-opus-4-8-highAnthropic12951288–130118,350
23Metamuse-spark-1.1Meta12941286–130111,255
24Ogpt-5.5-highOpenAI12921286–129825,488
25Ogpt-5.4OpenAI12921286–129922,828
26Anthropicclaude-sonnet-5.5-xhighAnthropic12901274–13051,608
27Anthropicclaude-opus-4-8Anthropic12891282–129618,857
28xgrok-4.5xAI12881280–129511,039
29Googlegemini-3-flashGoogle12851280–129139,142
30Ogpt-6.1-sol-maxOpenAI12851269–13011,408
31Anthropicclaude-sonnet-4-6Anthropic12831276–128927,429
32Moonshot AIkimi-k2.6Moonshot AI12821276–128916,387
33Alibabaqwen3.7-plusAlibaba12801273–128713,065
34Ogpt-6-astra-maxOpenAI12801268–12913,175
35Ogpt-5.6-sol-xhighOpenAI12791272–128710,421
36SStep 5 PreviewStepfun12781254–1302604
37Googlegemma-4-31bGoogle12771271–128339,111
38Anthropicclaude-sonnet-5-highAnthropic12751268–128213,029
39ByteDancedola-seed-2.0-proByteDance12751267–128210,981
40Ogpt-5.6-terra-xhighOpenAI12711264–127910,292
41Alibabaqwen3.8-27bAlibaba12701261–12796,422
42Moonshot AIkimi-k2.5-thinkingMoonshot AI12691263–127528,326
43Ogpt-5.2-chat-latest-20260210OpenAI12681261–127416,480
44Googlegemini-3.5-flash-liteGoogle12661258–12757,579
45Googlegemini-3-flash (thinking-minimal)Google12661260–127137,184
46xgrok-4.6-highxAI12651256–12745,736
47Zglm-5v-turboZ.ai12641259–126941,533
48Alibabaqwen3.5-397b-a17bAlibaba12641258–126931,441
49Xiaomimimo-v2.6-proXiaomi12641249–12781,787
50Googlegemini-2.5-proGoogle12631259–126888,813

Data: LMArena leaderboard dataset (CC BY 4.0). Changes: Top 50 entries per leaderboard; scores are rounded. https://huggingface.co/datasets/lmarena-ai/leaderboard-datasetCompany logos are trademarks of their respective owners and are used only for identification (icons: Simple Icons).Data: Epoch AI — Capabilities & benchmarking (CC BY 4.0). https://epoch.ai/benchmarks