Community battle rankings

LM Arena Leaderboard

Explore crowdsourced model rankings across different task dimensions based on LMSYS Chatbot Arena data.

Last updated Aug 18, 2026, 05:54 AMView official data

AGT

Agent Capability

Agent Capability - community model rankings
Rank	Model	Provider	Rating
1	Claude Opus 5 (High) Proprietary	anthropic	0.1219
2	Claude Fable 5 (High) Proprietary	anthropic	0.1201
3	Claude Opus 5 (Max) Proprietary	anthropic	0.1195
4	GPT 5.6 Sol (xHigh) Proprietary	openai	0.1086
5	Kimi K3 (Max) Kimi K3 license	moonshot	0.1060
6	Claude Opus 4.8 (High) Proprietary	anthropic	0.0978
7	GPT 5.5 (xHigh) Proprietary	openai	0.0890
8	Claude Opus 4.7 (High) Proprietary	anthropic	0.0817
9	GPT 5.5 (High) Proprietary	openai	0.0773
10	Claude Opus 4.7 Proprietary	anthropic	0.0766

Claude Opus 5 (High)

anthropic

0.1219

0 Votes

Claude Fable 5 (High)

anthropic

0.1201

0 Votes

Claude Opus 5 (Max)

anthropic

0.1195

0 Votes

GPT 5.6 Sol (xHigh)

openai

0.1086

0 Votes

Kimi K3 (Max)

moonshot

0.1060

0 Votes

Claude Opus 4.8 (High)

anthropic

0.0978

0 Votes

GPT 5.5 (xHigh)

openai

0.0890

0 Votes

Claude Opus 4.7 (High)

anthropic

0.0817

0 Votes

GPT 5.5 (High)

openai

0.0773

0 Votes

Claude Opus 4.7

anthropic

0.0766

0 Votes

DEV

Code & Web Development

Code & Web Development - community model rankings
Rank	Model	Provider	Rating	Votes
1	claude-opus-5-max Proprietary	anthropic	1691.5	6,448
2	kimi-k3-max Kimi K3 license	moonshot	1673.9	4,546
3	qwen3.8-max Proprietary	alibaba	1666.7	2,855
4	claude-opus-5-high Proprietary	anthropic	1662.6	7,334
5	grok-4.6-high Proprietary	xai	1631.1	1,513
6	claude-fable-5 Proprietary	anthropic	1626.9	7,867
7	gpt-5.6-sol-xhigh (codex-harness) Proprietary	openai	1621.6	8,595
8	gemini-3.7-flash-high Proprietary	google	1587.5	2,543
9	glm-5.2-max MIT	zai	1585.0	8,142
10	deepseek-v4-pro-high-20260813 MIT	deepseek	1584.5	2,180

claude-opus-5-max

anthropic

1691.5

6,448 Votes

kimi-k3-max

moonshot

1673.9

4,546 Votes

qwen3.8-max

alibaba

1666.7

2,855 Votes

claude-opus-5-high

anthropic

1662.6

7,334 Votes

grok-4.6-high

xai

1631.1

1,513 Votes

claude-fable-5

anthropic

1626.9

7,867 Votes

gpt-5.6-sol-xhigh (codex-harness)

openai

1621.6

8,595 Votes

gemini-3.7-flash-high

google

1587.5

2,543 Votes

glm-5.2-max

zai

1585.0

8,142 Votes

deepseek-v4-pro-high-20260813

deepseek

1584.5

2,180 Votes

DOC

Document Understanding

Document Understanding - community model rankings
Rank	Model	Provider	Rating	Votes
1	claude-fable-5 Proprietary	anthropic	1506.7	5,792
2	claude-opus-4-7 Proprietary	anthropic	1496.8	18,990
3	claude-opus-4-7-thinking Proprietary	anthropic	1495.8	18,793
4	claude-opus-4-6 Proprietary	anthropic	1494.7	37,271
5	claude-opus-4-6-thinking Proprietary	anthropic	1493.0	24,973
6	claude-opus-5-high Proprietary	anthropic	1492.7	1,663
7	claude-opus-4-8-thinking Proprietary	anthropic	1486.6	8,388
8	muse-spark-1.1 Proprietary	meta	1482.8	2,054
9	gpt-5.6-terra-xhigh Proprietary	openai	1482.3	1,969
10	claude-opus-4-8 Proprietary	anthropic	1481.8	8,193

claude-fable-5

anthropic

1506.7

5,792 Votes

claude-opus-4-7

anthropic

1496.8

18,990 Votes

claude-opus-4-7-thinking

anthropic

1495.8

18,793 Votes

claude-opus-4-6

anthropic

1494.7

37,271 Votes

claude-opus-4-6-thinking

anthropic

1493.0

24,973 Votes

claude-opus-5-high

anthropic

1492.7

1,663 Votes

claude-opus-4-8-thinking

anthropic

1486.6

8,388 Votes

muse-spark-1.1

meta

1482.8

2,054 Votes

gpt-5.6-terra-xhigh

openai

1482.3

1,969 Votes

claude-opus-4-8

anthropic

1481.8

8,193 Votes

EDIT

Image Editing

Image Editing - community model rankings
Rank	Model	Provider	Rating	Votes
1	gpt-image-2 (medium) Proprietary	openai	1462.7	184,189
2	muse-image Proprietary	meta	1404.7	47,573
3	mai-image-2.5 Proprietary	microsoft-ai	1401.9	155,115
4	seedream-5.0-pro Proprietary	bytedance	1392.6	71,266
5	chatgpt-image-latest-high-fidelity (20251216) Proprietary	openai	1389.8	494,514
6	grok-imagine-image-quality (20260519) Proprietary	xai	1389.7	35,613
7	gemini-3-pro-image-2k (nano-banana-pro) Proprietary	google	1389.1	501,853
8	gemini-3-pro-image-preview (nano-banana-pro) Proprietary	google	1385.3	519,141
9	gemini-3.1-flash-image (nano-banana-2) [web-search] Proprietary	google	1385.1	102,657
10	reve-2.1 Proprietary	reve	1374.5	19,093

gpt-image-2 (medium)

openai

1462.7

184,189 Votes

muse-image

meta

1404.7

47,573 Votes

mai-image-2.5

microsoft-ai

1401.9

155,115 Votes

seedream-5.0-pro

bytedance

1392.6

71,266 Votes

chatgpt-image-latest-high-fidelity (20251216)

openai

1389.8

494,514 Votes

grok-imagine-image-quality (20260519)

xai

1389.7

35,613 Votes

gemini-3-pro-image-2k (nano-banana-pro)

google

1389.1

501,853 Votes

gemini-3-pro-image-preview (nano-banana-pro)

google

1385.3

519,141 Votes

gemini-3.1-flash-image (nano-banana-2) [web-search]

google

1385.1

102,657 Votes

reve-2.1

reve

1374.5

19,093 Votes

I2C

Image to Code

Image to Code - community model rankings
Rank	Model	Provider	Rating	Votes
1	claude-opus-4-7-thinking Proprietary	Anthropic	1580.8	2,075
2	claude-sonnet-4-6 Proprietary	Anthropic	1556.6	3,158
3	claude-opus-4-7 Proprietary	Anthropic	1555.6	2,377
4	claude-opus-4-6-thinking Proprietary	Anthropic	1537.6	2,997
5	gpt-5.5-xhigh (codex-harness) Proprietary	OpenAI	1536.8	1,816
6	claude-opus-4-6 Proprietary	Anthropic	1534.3	3,043
7	kimi-k2.6 Modified MIT	Moonshot	1522.3	1,451
8	gpt-5.5-high (codex-harness) Proprietary	OpenAI	1519.1	1,965
9	gemini-3.1-pro-preview Proprietary	Google	1489.9	3,597
10	gpt-5.5 (codex-harness) Proprietary	OpenAI	1488.7	1,935

claude-opus-4-7-thinking

Anthropic

1580.8

2,075 Votes

claude-sonnet-4-6

Anthropic

1556.6

3,158 Votes

claude-opus-4-7

Anthropic

1555.6

2,377 Votes

claude-opus-4-6-thinking

Anthropic

1537.6

2,997 Votes

gpt-5.5-xhigh (codex-harness)

OpenAI

1536.8

1,816 Votes

claude-opus-4-6

Anthropic

1534.3

3,043 Votes

kimi-k2.6

Moonshot

1522.3

1,451 Votes

gpt-5.5-high (codex-harness)

OpenAI

1519.1

1,965 Votes

gemini-3.1-pro-preview

Google

1489.9

3,597 Votes

gpt-5.5 (codex-harness)

OpenAI

1488.7

1,935 Votes

I2V

Image to Video

Image to Video - community model rankings
Rank	Model	Provider	Rating	Votes
1	minimax-h3 minimax-h3-community-license-agreement	minimax	1489.4	13,020
2	dreamina-seedance-2.5-720p Proprietary	bytedance	1483.7	2,912
3	dreamina-seedance-2.0-720p Proprietary	bytedance	1478.2	105,298
4	gemini-omni-flash Proprietary	google	1462.2	54,179
5	grok-imagine-video-1.5-720p Proprietary	xai	1460.3	99,649
6	flux-3-video-20260811 Proprietary	bfl	1449.3	3,770
7	happyhorse-1.0 Proprietary	aorizon	1441.5	70,772
8	wan2.7-i2v Proprietary	wan	1427.5	57,878
9	grok-imagine-video-720p Proprietary	xai	1415.0	517,108
10	veo-3.1-audio Proprietary	google	1397.3	25,156

minimax-h3

minimax

1489.4

13,020 Votes

dreamina-seedance-2.5-720p

bytedance

1483.7

2,912 Votes

dreamina-seedance-2.0-720p

bytedance

1478.2

105,298 Votes

gemini-omni-flash

google

1462.2

54,179 Votes

grok-imagine-video-1.5-720p

xai

1460.3

99,649 Votes

flux-3-video-20260811

bfl

1449.3

3,770 Votes

happyhorse-1.0

aorizon

1441.5

70,772 Votes

wan2.7-i2v

wan

1427.5

57,878 Votes

grok-imagine-video-720p

xai

1415.0

517,108 Votes

veo-3.1-audio

google

1397.3

25,156 Votes

WEB

Web Search

Web Search - community model rankings
Rank	Model	Provider	Rating	Votes
1	claude-fable-5 Proprietary	anthropic	1236.8	20,630
2	gpt-5.5-search Proprietary	openai	1223.5	67,573
3	claude-opus-4-6-search Proprietary	anthropic	1222.7	112,201
4	claude-opus-4-7 Proprietary	anthropic	1211.0	68,515
5	gemini-3.1-pro-grounding Proprietary	google	1209.7	89,973
6	grok-4.20-multi-agent-beta-0309 Proprietary	xai	1206.8	87,118
7	gemini-3-pro-grounding Proprietary	google	1201.2	37,255
8	claude-sonnet-4-6-search Proprietary	anthropic	1199.6	111,959
9	grok-4.20-beta1 Proprietary	xai	1197.5	53,877
10	claude-opus-4-8 Proprietary	anthropic	1195.6	48,610

claude-fable-5

anthropic

1236.8

20,630 Votes

gpt-5.5-search

openai

1223.5

67,573 Votes

claude-opus-4-6-search

anthropic

1222.7

112,201 Votes

claude-opus-4-7

anthropic

1211.0

68,515 Votes

gemini-3.1-pro-grounding

google

1209.7

89,973 Votes

grok-4.20-multi-agent-beta-0309

xai

1206.8

87,118 Votes

gemini-3-pro-grounding

google

1201.2

37,255 Votes

claude-sonnet-4-6-search

anthropic

1199.6

111,959 Votes

grok-4.20-beta1

xai

1197.5

53,877 Votes

claude-opus-4-8

anthropic

1195.6

48,610 Votes

TXT

Text Chat

Text Chat - community model rankings
Rank	Model	Provider	Rating	Votes
1	claude-fable-5 Proprietary	anthropic	1506.6	21,439
2	claude-opus-4-6-high Proprietary	anthropic	1504.7	72,466
3	claude-opus-4-7-high Proprietary	anthropic	1502.2	60,268
4	muse-spark-1.2 (xHigh) Proprietary	meta	1498.7	3,259
5	claude-opus-4-6 Proprietary	anthropic	1497.2	76,428
6	claude-opus-4-7 Proprietary	anthropic	1494.1	61,369
7	claude-opus-5-high Proprietary	anthropic	1493.7	19,792
8	claude-opus-5-max Proprietary	anthropic	1491.1	9,559
9	qwen3.8-max Proprietary	alibaba	1490.7	6,908
10	muse-spark-1.1 Proprietary	meta	1489.0	16,805

claude-fable-5

anthropic

1506.6

21,439 Votes

claude-opus-4-6-high

anthropic

1504.7

72,466 Votes

claude-opus-4-7-high

anthropic

1502.2

60,268 Votes

muse-spark-1.2 (xHigh)

meta

1498.7

3,259 Votes

claude-opus-4-6

anthropic

1497.2

76,428 Votes

claude-opus-4-7

anthropic

1494.1

61,369 Votes

claude-opus-5-high

anthropic

1493.7

19,792 Votes

claude-opus-5-max

anthropic

1491.1

9,559 Votes

qwen3.8-max

alibaba

1490.7

6,908 Votes

muse-spark-1.1

meta

1489.0

16,805 Votes

T2I

Text to Image

Text to Image - community model rankings
Rank	Model	Provider	Rating	Votes
1	gpt-image-2 (medium) Proprietary	openai	1380.5	69,194
2	reve-2.1 Proprietary	reve	1301.6	7,598
3	muse-image Proprietary	meta	1283.4	14,510
4	reve-2.0 Proprietary	reve	1270.1	14,650
5	gemini-3.1-flash-image (nano-banana-2) [web-search] Proprietary	google	1263.1	27,910
6	qwen-image-3.0-pro Proprietary	alibaba	1257.7	3,735
7	seedream-5.0-pro Proprietary	bytedance	1256.9	26,510
8	mai-image-2.5 Proprietary	microsoft-ai	1256.1	44,940
9	gemini-3.1-flash-lite-image (nano-banana-2-lite) Proprietary	google	1250.9	16,419
10	gemini-3-pro-image-2k (nano-banana-pro) Proprietary	google	1245.6	138,756

gpt-image-2 (medium)

openai

1380.5

69,194 Votes

reve-2.1

reve

1301.6

7,598 Votes

muse-image

meta

1283.4

14,510 Votes

reve-2.0

reve

1270.1

14,650 Votes

gemini-3.1-flash-image (nano-banana-2) [web-search]

google

1263.1

27,910 Votes

qwen-image-3.0-pro

alibaba

1257.7

3,735 Votes

seedream-5.0-pro

bytedance

1256.9

26,510 Votes

mai-image-2.5

microsoft-ai

1256.1

44,940 Votes

gemini-3.1-flash-lite-image (nano-banana-2-lite)

google

1250.9

16,419 Votes

gemini-3-pro-image-2k (nano-banana-pro)

google

1245.6

138,756 Votes

T2V

Text to Video

Text to Video - community model rankings
Rank	Model	Provider	Rating	Votes
1	gemini-omni-flash Proprietary	google	1511.6	16,916
2	flux-3-video Proprietary	bfl	1494.4	1,291
3	dreamina-seedance-2.0-720p Proprietary	bytedance	1481.7	49,058
4	dreamina-seedance-2.5-720p Proprietary	bytedance	1477.0	1,337
5	muse-video Proprietary	meta	1457.1	2,176
6	minimax-h3 minimax-h3-community-license-agreement	minimax	1452.7	3,081
7	happyhorse-1.0 Proprietary	aorizon	1428.0	22,113
8	sora-2-pro Proprietary	openai	1364.0	46,509
9	veo-3.1-audio Proprietary	google	1363.9	13,743
10	veo-3.1-audio-1080p Proprietary	google	1362.9	24,979

gemini-omni-flash

google

1511.6

16,916 Votes

flux-3-video

bfl

1494.4

1,291 Votes

dreamina-seedance-2.0-720p

bytedance

1481.7

49,058 Votes

dreamina-seedance-2.5-720p

bytedance

1477.0

1,337 Votes

muse-video

meta

1457.1

2,176 Votes

minimax-h3

minimax

1452.7

3,081 Votes

happyhorse-1.0

aorizon

1428.0

22,113 Votes

sora-2-pro

openai

1364.0

46,509 Votes

veo-3.1-audio

google

1363.9

13,743 Votes

veo-3.1-audio-1080p

google

1362.9

24,979 Votes

V2V

Video Editing

Video Editing - community model rankings
Rank	Model	Provider	Rating	Votes
1	dreamina-seedance-2.5-720p Proprietary	bytedance	1410.8	403
2	minimax-h3 minimax-h3-community-license-agreement	minimax	1388.2	554
3	dreamina-seedance-2.0-720p Proprietary	bytedance	1359.5	3,634
4	gemini-omni-flash Proprietary	google	1357.9	1,868
5	happyhorse-1.0 Proprietary	aorizon	1307.3	3,115
6	grok-imagine-video Proprietary	xai	1262.0	13,281
7	kling-o3-pro Proprietary	kling	1256.5	7,125
8	kling-o1-pro Proprietary	kling	1199.9	10,104
9	runway-gen4-aleph Proprietary	runway	1182.9	9,817

dreamina-seedance-2.5-720p

bytedance

1410.8

403 Votes

minimax-h3

minimax

1388.2

554 Votes

dreamina-seedance-2.0-720p

bytedance

1359.5

3,634 Votes

gemini-omni-flash

google

1357.9

1,868 Votes

happyhorse-1.0

aorizon

1307.3

3,115 Votes

grok-imagine-video

xai

1262.0

13,281 Votes

kling-o3-pro

kling

1256.5

7,125 Votes

kling-o1-pro

kling

1199.9

10,104 Votes

runway-gen4-aleph

runway

1182.9

9,817 Votes

VIS

Vision Understanding

Vision Understanding - community model rankings
Rank	Model	Provider	Rating	Votes
1	claude-fable-5 Proprietary	anthropic	1314.7	8,333
2	qwen3.8-max Proprietary	alibaba	1301.3	5,344
3	claude-opus-4-7-thinking Proprietary	anthropic	1301.1	20,461
4	claude-opus-4-6-thinking Proprietary	anthropic	1300.1	20,310
5	claude-opus-4-7 Proprietary	anthropic	1299.0	20,857
6	claude-opus-5-high Proprietary	anthropic	1296.8	3,821
7	gemini-3.6-flash Proprietary	google	1294.7	1,205
8	muse-spark Proprietary	meta	1294.2	5,576
9	claude-opus-4-6 Proprietary	anthropic	1292.9	24,609
10	muse-spark-1.2 (xHigh) Proprietary	meta	1290.0	1,282

claude-fable-5

anthropic

1314.7

8,333 Votes

qwen3.8-max

alibaba

1301.3

5,344 Votes

claude-opus-4-7-thinking

anthropic

1301.1

20,461 Votes

claude-opus-4-6-thinking

anthropic

1300.1

20,310 Votes

claude-opus-4-7

anthropic

1299.0

20,857 Votes

claude-opus-5-high

anthropic

1296.8

3,821 Votes

gemini-3.6-flash

google

1294.7

1,205 Votes

muse-spark

meta

1294.2

5,576 Votes

claude-opus-4-6

anthropic

1292.9

24,609 Votes

muse-spark-1.2 (xHigh)

meta

1290.0

1,282 Votes

How to read the leaderboard

Data source: LMSYS Chatbot Arena

Rankings are based on anonymous, head-to-head model battles voted on by real users. Rating scales can differ between leaderboards, so scores are best compared within the same leaderboard rather than across tasks.

Rating

A relative capability score calculated from community battle outcomes. Higher values indicate stronger performance within the current leaderboard.

Votes

The number of valid community votes collected through anonymous model battles, indicating the evidence behind each rating.