← Arena leaderboards

Text · Factuality · Polish

Arena · 50 models

RankModelScore
1Gemini 3 Pro1,487.9749
2Claude Opus 5 (Max)1,484.2527
3Claude Opus 4.6 (High)1,481.913
4GPT-5.5 (High)1,477.6909
5GPT-5.4 (High)1,476.2388
6Gemini 3 Flash1,476.2086
7Gemini 3.1 Pro Preview1,475.2182
8Claude Opus 5 (High)1,470.8018
9Gemini 3.5 Flash (Medium)1,468.1076
10GPT-5.6 Sol (xHigh)1,467.6778
11GPT-5.51,467.1958
12Claude Opus 4.71,465.4845
13Claude Opus 4.61,465.0807
14Claude Opus 4.7 (High)1,463.639
15Gemini 3.8 Flash (High)1,458.4708
16Claude Opus 4.8 (High)1,458.0061
17Claude Opus 4.81,456.8375
18GPT-5.41,456.5456
19Gemini 2.5 Pro1,455.5588
20Gemini 3 Flash (Minimal)1,454.1403
21MiMo-V2.5-Pro1,451.3257
22GPT-5.6 Terra (xHigh)1,448.8407
23Grok 4.51,448.5991
24GPT-5.11,445.5435
25ChatGPT-4o Latest (20250326)1,444.5582
26GPT-5.21,441.226
27Dola Seed 2.0 Pro1,440.695
28GLM-5.11,440.4333
29Qwen3.5-397B-A17B1,439.4208
30Grok 4.20 Multi-Agent Beta (0309)1,439.3757
31GPT-5.1 (High)1,437.9658
32DeepSeek-V4-Pro Preview (High)1,437.6445
33Claude Opus 4.5 (20251101)1,437.4251
34Qwen3.7-Plus1,437.1314
35GPT-5.4 mini (High)1,435.225
36Gemini 3.1 Flash-Lite Preview1,432.55
37Claude Opus 4.5 (High, 32K, 20251101)1,431.4987
38Grok 4.20 Beta (Reasoning, 0309)1,428.8455
39ERNIE 5.0 (0110)1,427.9019
40Claude Sonnet 4.61,426.7642
41MiniMax-M31,426.0033
42Grok 4.1 (Thinking)1,425.7662
43GPT-5.2 (High)1,425.6735
44GLM-5.3-Flash1,425.3849
45Kimi K2.5 (Thinking)1,425.3667
46GLM-5.2 (Max)1,422.6795
47Grok 4.11,420.6306
48Mistral Large 31,416.7685
49GLM-4.61,414.4021
50Gemini 2.5 Flash1,409.6292