← Arena leaderboards

Text · Style Control · Korean

Arena · 50 models

RankModelScore
1Claude Fable 5 (High)1,496.6257
2Claude Opus 5 (Max)1,494.6864
3Claude Fable 5.1 (Max)1,494.4979
4Claude Opus 5 (High)1,493.1174
5Muse Spark 1.3 (Max)1,491.2854
6Gemini 3.7 Flash (High)1,480.5543
7Muse Spark1,473.6478
8Muse Spark 1.11,473.6299
9Claude Opus 4.71,470.6259
10Kimi K3 (Max)1,470.0493
11GPT-6.1 Sol (Max)1,466.5882
12Gemini 3.8 Flash (High)1,465.4496
13DeepSeek-V4-Pro (High, 20260813)1,464.3749
14Muse Spark 1.2 (xHigh)1,463.0013
15Claude Opus 4.61,462.6502
16Claude Opus 4.7 (High)1,461.4357
17GLM-5.3 (Max)1,460.1085
18Qwen3.8-Max1,456.4023
19Gemini 3.1 Pro Preview1,454.968
20GPT-5.6 Sol (xHigh)1,454.6493
21GPT-6 Astra (Max)1,453.2452
22GPT-5.5 Instant1,450.6546
23Gemini 3 Pro1,448.9648
24Claude Opus 4.5 (20251101)1,447.7892
25Claude Opus 4.8 (High)1,447.7528
26Claude Opus 4.6 (High)1,446.7334
27Gemini 3 Flash1,444.8837
28DeepSeek-V4.1-Flash-20260910 (Max)1,444.4948
29GPT-5.5 (High)1,443.9119
30Grok 4.20 Beta 11,443.8806
31GLM-5.3-Flash1,442.9083
32GLM-5.2 (Max)1,440.1083
33Gemini 3.5 Flash (Medium)1,439.4943
34Claude Opus 4.81,439.1239
35MiMo-V2.5-Pro1,436.7366
36Gemini 3.5 Flash (High)1,435.5198
37DeepSeek-V4-Pro Preview (High)1,434.9816
38Grok 4.20 Multi-Agent Beta (0309)1,433.607
39GPT-5.4 (High)1,433.0403
40Claude Sonnet 5 (High)1,432.7288
41Grok 4.20 Beta (Reasoning, 0309)1,432.6767
42Claude Sonnet 4.61,431.6375
43Grok 4.11,431.5897
44GPT-5.51,430.3211
45MiMo-V2.6-Flash1,429.2017
46Gemini 3 Flash (Minimal)1,429.1086
47Gemini 3.6 Flash (High)1,428.1935
48GLM-51,426.6591
49GPT-5.41,426.444
50Grok 4.1 (Thinking)1,426.4206