排行榜总览 (截止 2026年7月)

这里汇总主流 AI 模型在文本、图像、视觉理解等不同 Arena 中的表现,方便快速查看各类任务下的领先模型。 查看官方榜单。

第一名
第二名
第三名
模型 Overall专家评测高难提示词代码数学创意写作指令遵循长查询
claude-fable-5 11111111
claude-opus-4-6 22232212
claude-opus-4-7 43345544
claude-opus-4-6-thinking 35414334
claude-opus-4-7-thinking 33523743
muse-spark 615101130132431
gemini-3.1-pro-preview 7101115176911
gemini-3-pro 81918181741410
claude-opus-4-8-thinking 986762057
gpt-5.5-high 10111210733119
claude-opus-4-6 11778112286
gpt-5.4-high 124859101213
gemini-3.1-flash 131534343301333
gpt-4.20-bot-latest 1427203231293024

Agent

1
Claude Fable 5 (High)
AI
+ 13.34%
+1.55%
2
Claude Opus 4.8 (Thinking)
AI
+ 9.37%
+1.29%
3
GPT 5.5 (xHigh)
+ 8.21%
+1.02%
4
Claude Opus 4.7
AI
+ 8.16%
+1.28%
5
Claude Opus 4.7 (Thinking)
AI
+ 8.07%
+1.23%
6
GPT 5.5 (High)
+ 7.13%
+0.79%
7
GLM 5.2 (Max)
Z
+ 6.93%
+0.40%
8
GPT 5.4 (High)
+ 6.65%
+0.75%
9
Claude Opus 4.6
AI
+ 6.47%
+1.21%
10
GPT 5.5
+ 6.22%
+0.72%

Text

Overall
1
claude-fable-5
AI
1508
±9
2
claude-opus-4-6-thinking
AI
1503
±4
3
claude-opus-4-7-thinking
AI
1502
±4
4
claude-opus-4-6
AI
1499
±4
5
claude-opus-4-7
AI
1494
±4
6
muse-spark
1487
±4
7
gemini-3.1-pro-preview
G
1486
±4
8
gemini-3-pro
G
1486
±4
9
claude-opus-4-8-thinking
AI
1484
±4
10
gpt-5.5-high
1481
±5

WebDev

Overall
1
claude-fable-5
AI
1654
+16/-16
2
glm-5.2 (max)
Z
1593
+15/-15
3
claude-opus-4-8-thinking
AI
1565
+12/-12
4
claude-opus-4-7-thinking
AI
1563
+8/-8
5
claude-opus-4-7
AI
1557
+8/-8
6
claude-opus-4-8
AI
1542
+14/-11
7
claude-opus-4-6-thinking
AI
1542
+15/-12
8
seed-2.1-pro-preview
1539
+15/-15
9
claude-opus-4-6
AI
1538
+11/-11
10
qwen3.7-max-20260517
Z
1530
+14/-13

Vision

Overall
1
claude-fable-5
AI
1311
±14
2
claude-opus-4-7-thinking
AI
1308
±7
3
claude-opus-4-6-thinking
AI
1299
±7
4
claude-opus-4-7
AI
1298
±7
5
claude-opus-4-6
AI
1297
±7
6
muse-spark
1294
±10
7
claude-opus-4-8-thinking
AI
1290
±8
8
gemini-3-pro
G
1289
±8
9
gpt-5.5
1285
±8
10
gpt-5.4-high
1283
±7

Document

1
claude-opus-4-6
AI
1507
±6
2
claude-opus-4-6-thinking
AI
1507
±7
3
claude-opus-4-7-thinking
AI
1498
±7
4
claude-opus-4-7
AI
1496
±7
5
claude-fable-5
AI
1486
±7
6
claude-sonnet-4-6
AI
1487
±8
7
gpt-5.5-high
1485
±7
8
gpt-5.5
1483
±7
9
gpt-5.4
1474
±8
10
claude-opus-4-8-thinking
AI
1473
±13

Text-to-Image

1
gpt-image-2 (medium)
1387
±15
2
reve 2.0
R
1273
±17
3
gemini-3.1-flash-image-preview
G
1270
±22
4
mai-image-2.5
M
1257
±17
5
gemini-3.1-flash-lite-image
G
1251
±18
6
gemini-3-pro-image-preview
G
1245
±14
7
gpt-image-1.5-high-fidelity
1241
±11
8
gemini-3-pro-image-preview
G
1232
±13
9
grok-imagine-image-quality
x1
1231
±12
10
ideogram-4.0-quality
R
1209
±10

Image Edit

Single Image Edit
1
gpt-image-2 (medium)
1464
±4
2
mai-image-2.5
M
1403
±6
3
chatgpt-image-latest-high-fidelity
1390
±8
4
grok-imagine-image-quality
x1
1389
±8
5
gemini-3-pro-image-preview
G
1388
±8
6
gemini-3.1-flash-image-preview
G
1387
±6
7
gemini-3-pro-image-preview
G
1385
±8
8
gpt-image-1.5-high-fidelity
1373
±5
9
grok-imagine-image-quality
x1
1358
±8
10
reve 2.0
R
1357
±7

Image-to-WebDev

1
claude-opus-4-7-thinking
AI
1581
+15/-15
2
claude-sonnet-4-6
AI
1567
+15/-15
3
claude-opus-4-7
AI
1556
+12/-14
4
claude-opus-4-6-thinking
AI
1538
+13/-13
5
gpt-5.5-high (code-names)
1537
+15/-15
6
claude-opus-4-6
AI
1534
+13/-13
7
kimi-k2.6
K
1522
+17/-17
8
gpt-5.5-high (code-names)
1519
+15/-15
9
gemini-3.1-pro-preview
G
1490
+19/-12
10
gpt-5.5 (code-names)
1480
+15/-15

Search

1
claude-opus-4-6-search
AI
1252
±10
2
gpt-5.5-search
1240
±12
3
claude-fable-5
AI
1237
±11
4
claude-opus-4-7
AI
1232
±12
5
ernie-5.1
DS
1226
±10
6
claude-sonnet-4-6-search
AI
1218
±9
7
gemini-3.1-pro-grounding
G
1214
±10
8
gemini-3-pro-grounding
G
1207
±10
9
grok-4.20-multi-agent-beta-0309
x1
1206
±9
10
gpt-5.2-search
1205
±10

Text-to-Video

1
gemini-omni-flash
G
1527
±7
2
dreamina-seedance-2.0-720p
1468
±7
3
happyhorse-1.0
R
1437
±7
4
veo-3.1-audio-1080p
G
1369
±11
5
wan2.2-t2v
G
1368
±13
6
veo-3.1-audio
G
1365
±14
7
sora-2-pro
1365
±8
8
veo-3.1-fast-audio
G
1364
±11
9
veo-3.1-fast-audio-1080p
G
1362
±11
10
grok-imagine-video-720p
x1
1358
±8

Image-to-Video

1
dreamina-seedance-2.0-720p
1474
±10
2
gemini-omni-flash
G
1469
±11
3
grok-imagine-video-1.5-preview-720p
x1
1466
±5
4
happyhorse-1.0
R
1444
±10
5
wan2.2-i2v
G
1424
±12
6
grok-imagine-video-720p
x1
1422
±8
7
veo-3.1-studio
G
1398
±11
8
veo-3.1-audio-1080p
G
1391
±8
9
veo-3.1-fast-audio
G
1385
±8
10
grok-imagine-video-480p
x1
1384
±8

Video Edit

1
dreamina-seedance-2.0-720p
1377
±12
2
gemini-omni-flash
G
1347
±16
3
happyhorse-1.0
R
1308
±12
4
grok-imagine-video
x1
1284
±11
5
ling-03-pro
1251
±12
6
ling-03-pro
1203
±8
7
runway-gen4-aleph
R
1194
±8