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
1Gemini 4 CarbonHigh· LMArena data1,547.3
2Gemini 4 ArgonHigh· LMArena data1,537.7
3Claude Opus 5.5High· LMArena data1,534.4
4Claude Opus 4.6High· LMArena data1,530.1
5Claude Opus 4.6· LMArena data1,528.3
6Gemini 4 NeonHigh· LMArena data1,524.4
7Claude Fable 5.1Max· LMArena data1,523.3
8Claude Opus 5Max· LMArena data1,522.7
9Muse WonderMax· LMArena data1,518.8
10Claude Opus 5High· LMArena data1,518.4
11MiMo V2.6 Pro· LMArena data1,516.6
12Gemini 4 HydrogenHigh· LMArena data1,511.2
13Claude Fable 5High· LMArena data1,510.4
14Claude Opus 4.7High· LMArena data1,507.8
15Gemini 3.8 FlashHigh· LMArena data1,505.5
16Qwen3.8 Max· LMArena data1,504.7
17Claude Opus 4.7· LMArena data1,501.7
18Muse Spark 1.3Max· LMArena data1,500.2
19GLM 5.3Max· LMArena data1,499.9
20Claude Sonnet 5.5xHigh· LMArena data1,494.2
21MiMo V2.5 Pro· LMArena data1,494
22Claude Sonnet 4.6· LMArena data1,493.6
23Kimi K3Max· LMArena data1,492.6
24Muse Spark 1.2xHigh· LMArena data1,492.5
25GLM 5.3 Flash· LMArena data1,491.1
26Gemini 3.7 FlashHigh· LMArena data1,488.9
27MiMo V2.6 Flash· LMArena data1,488.5
28DeepSeek V4.1 Flash 0910Max· LMArena data1,487.4
29Claude Opus 4.8High· LMArena data1,486.9
30ERNIE 5.1· LMArena data1,484.9
31Kimi K3.1 ProMax· LMArena data1,483.6
32Gemini 3.6 FlashHigh· LMArena data1,483.4
33Qwen3.5-Max-Preview· LMArena data1,483.1
34GPT 5.5· LMArena data1,483.1
35Gemini 3.5 FlashHigh· LMArena data1,482.9
36Claude Fable 5.2Max· LMArena data1,482.8
37GPT 5.5High· LMArena data1,482.7
38Claude Opus 4.520251101· LMArena data1,482.3
39GLM 5.1· LMArena data1,482
40GLM 5.2Max· LMArena data1,481.4
41Muse Spark 1.1· LMArena data1,481.2
42Gemini 3.1 Pro Preview· LMArena data1,480
43GPT 5.4High· LMArena data1,479.2
44Gemini 3.5 FlashMedium· LMArena data1,479
45GPT-5.6 SolxHigh· LMArena data1,478.6
46Muse Spark· LMArena data1,477.5
47Claude Opus 4.5High, 32K, 20251101· LMArena data1,477.2
48Claude Opus 4.8· LMArena data1,475.1
49Gemini 3 Pro Preview· LMArena data1,474.6
50MiMo V2 Pro· LMArena data1,473.1

Showing 1–50 of 421 results.

Data credits: LMArena, CC BY 4.0 (licence)