← Arena leaderboards

Text · Polish

Arena · 50 models

RankModelScore
1Gemini 4 Argon (High)1,537.5168
2Claude Fable 5.1 (Max)1,535.7577
3Claude Opus 5 (Max)1,521.2133
4Claude Opus 5 (High)1,520.889
5Gemini 3.5 Flash (High)1,517.1376
6Gemini 3.7 Flash (High)1,504.2524
7Gemini 3 Pro1,502.7129
8Gemini 3.8 Flash (High)1,499.4628
9Gemini 3.5 Flash (Medium)1,498.5062
10Claude Opus 4.6 (High)1,497.9628
11Muse Spark 1.11,497.2022
12Gemini 3.1 Pro Preview1,496.9342
13Gemini 3 Flash1,493.9822
14Claude Fable 5 (High)1,492.7302
15Muse Spark 1.3 (Max)1,491.4598
16Claude Opus 4.71,489.9538
17Claude Opus 4.61,488.5356
18Dola Seed 2.0 Pro1,486.9677
19ERNIE-5.0-Preview (1022)1,482.9185
20ERNIE-5.0-Preview (1203)1,482.7211
21Claude Opus 4.7 (High)1,479.0479
22Gemini 2.5 Pro1,478.9378
23Qwen3.8-Max1,475.3096
24Gemini 3.6 Flash (High)1,474.8036
25Muse Spark1,474.5531
26GPT-5.4 (High)1,472.4115
27Gemini 3 Flash (Minimal)1,468.4288
28ERNIE 5.11,467.8934
29ERNIE 5.0 (0110)1,467.1491
30GPT-5.5 (High)1,465.8302
31MiMo-V2.5-Pro1,463.455
32Kimi K3 (Max)1,462.1219
33GLM-5.3 (Max)1,461.6442
34DeepSeek-V4-Pro1,460.7475
35GPT-5.51,459.5678
36Claude Opus 4.81,458.3781
37DeepSeek-V4.1-Flash-20260910 (Max)1,457.6244
38Claude Opus 4.8 (High)1,457.1474
39GLM-5.2 (Max)1,456.6496
40Grok 4.1 (Thinking)1,454.9226
41GPT-5.41,454.7126
42DeepSeek-V4-Pro Preview (High)1,454.1317
43Qwen3.5-Max-Preview1,454.1282
44Grok 4.20 Multi-Agent Beta (0309)1,453.9901
45GPT-5.6 Sol (xHigh)1,452.9885
46GLM-5.11,451.8301
47GLM-5.3-Flash1,451.7704
48Grok 4.11,451.5936
49Inkling1,451.5127
50Grok 4.20 Beta (Reasoning, 0309)1,451.4406