← Arena leaderboards

Text · Hard Prompts

Arena · 50 models

RankModelScore
1Gemini 4 Carbon (High)1,552.5086
2Gemini 4 Argon (High)1,541.7328
3Claude Opus 5.5 (High)1,536.0991
4Claude Fable 5.1 (Max)1,528.9959
5Claude Opus 5 (Max)1,526.4038
6Claude Opus 4.6 (High)1,524.6485
7Claude Opus 5 (High)1,523.3298
8Claude Opus 4.61,521.669
9MiMo-V2.6-Pro1,511.1768
10Gemini 3.8 Flash (High)1,508.282
11Claude Fable 5 (High)1,507.6001
12Muse Wonder (Max)1,506.9383
13Claude Opus 4.7 (High)1,505.6477
14Muse Spark 1.3 (Max)1,502.5362
15Claude Opus 4.71,499.5271
16Gemini 3.7 Flash (High)1,497.2772
17Qwen3.8-Max1,496.3632
18Kimi K3 (Max)1,495.5364
19Claude Sonnet 5.5 (xHigh)1,493.7649
20GLM-5.3-Flash1,492.2255
21Muse Spark 1.2 (xHigh)1,491.6424
22GLM-5.3 (Max)1,490.2866
23GPT-5.5 (High)1,488.9092
24Gemini 3.5 Flash (High)1,488.1939
25Muse Spark 1.11,485.8665
26MiMo-V2.5-Pro1,485.8056
27GPT-5.51,485.7869
28Gemini 3.6 Flash (High)1,485.6751
29Gemini 3.5 Flash (Medium)1,485.1641
30GPT-5.4 (High)1,484.9891
31Gemini 3.1 Pro Preview1,484.5649
32Claude Sonnet 4.61,484.3056
33GPT-5.6 Sol (xHigh)1,483.1449
34DeepSeek-V4.1-Flash-20260910 (Max)1,482.7904
35Qwen3.5-Max-Preview1,482.7424
36Claude Opus 4.8 (High)1,482.2205
37Qwen3.7-Max-Preview1,481.8714
38ERNIE 5.11,480.6365
39GLM-5.2 (Max)1,480.5632
40Gemini 3 Pro1,480.2621
41Claude Opus 4.5 (20251101)1,475.9836
42Muse Spark1,474.2018
43Claude Opus 4.81,474.1412
44MiMo-V2.6-Flash1,473.3554
45Claude Opus 4.5 (High, 32K, 20251101)1,473.2681
46GLM-5.11,473.0448
47Kimi K2.61,470.5389
48GPT-5.6 Terra (xHigh)1,468.9819
49GPT-5.41,468.6528
50GPT-6.1 Sol (Max)1,467.4611