Skip to content

Benchmark

Arena ELO - Hard Prompts English

People compare model answers and choose which they prefer. Higher scores are better.

Models measured
364
Scores recorded
421
Sources
1
IQ weight
1.31

Score distribution

1,319.1
Mean
1,374
Median
157.7
Std dev
1,085.3
10th percentile
1,479.2
90th percentile
734.7
Min
734.71,547.3
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.31 for chat models, measured over 364 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51GPT-6.1 SolMax· LMArena data1,472.6
52Qwen3.7 Max Preview· LMArena data1,472.5
53Kimi K2.6· LMArena data1,471.4
54Claude Sonnet 4.5High, 32K, 20250929· LMArena data1,471
55Claude Sonnet 4.520250929· LMArena data1,470.3
56GPT-6 AstraMax· LMArena data1,467.7
57DeepSeek V4 Pro· LMArena data1,466.6
58GPT-5.6 TerraxHigh· LMArena data1,465
59Claude Sonnet 5High· LMArena data1,464.7
60Step 5 Preview· LMArena data1,464.3
61GPT 5.4· LMArena data1,464
62Grok 4.5· LMArena data1,464
63Qwen 3.8 27B· LMArena data1,462.9
64nova-experimental-chat26-02-10· LMArena data1,462.9
65DeepSeek V4 ProHigh, 20260813· LMArena data1,462.6
66GPT 5.1High· LMArena data1,462.2
67Gemini 3 Flash· LMArena data1,461.8
68MiMo V2.5· LMArena data1,460.8
69Kimi K2.5Thinking· LMArena data1,460.4
70DeepSeek-V4-Pro PreviewHigh· LMArena data1,460
71GLM 5· LMArena data1,459.9
72Qwen3.7 Plus· LMArena data1,459.6
73GPT-5.6 LunaxHigh· LMArena data1,459.4
74LongCat-Flash-Chat Exp2602· LMArena data1,458.7
75Qwen3.6 Max Preview· LMArena data1,455.8
76MiMo V2 Omni· LMArena data1,455.7
77Gemma 4 31B IT· LMArena data1,455.6
78Qwen3.6 Plus· LMArena data1,455.3
79Grok 4.20Reasoning, 0309, beta· LMArena data1,455
80Beam 1.0 501B A23B· LMArena data1,454.4
81Hy3 (Deep Infra)· LMArena data1,452.5
82GPT-6.1 TerraMax· LMArena data1,452.4
83Gemini 2.5 Pro· LMArena data1,452.3
84GLM 4.7· LMArena data1,452
85MiniMax M3· LMArena data1,451.9
86Nemotron 3 UltraNVFP4· LMArena data1,451.8
87Claude Opus 4.1Thinking 16K, 20250805· LMArena data1,451
88GPT-5.2 Chat20260210· LMArena data1,450.5
89Inkling FP4· LMArena data1,450.5
90Kimi K2.5Instant· LMArena data1,450.5
91Doubao Seed 2.0 Pro· LMArena data1,450.4
92Qwen3.5 397B A17B· LMArena data1,448.9
93Gemma 4 26B A4B· LMArena data1,448.6
94DeepSeek V4 Flash· LMArena data1,448.4
95Qwen3 Max Preview· LMArena data1,448.1
96ERNIE 5.00110· LMArena data1,448
97Claude Opus 4.120250805· LMArena data1,447.9
98nova-experimental-chat12-10· LMArena data1,447
99Grok 4.20 Multi-Agent0309, beta· LMArena data1,446.6
100DeepSeek V3.2Thinking· LMArena data1,445.6

Showing 51–100 of 421 results.

Data credits: LMArena, CC BY 4.0 (licence)