Skip to content

Benchmark

Arena ELO - Multi Turn

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

Models measured
364
Scores recorded
421
Sources
1
IQ weight
1.21

Score distribution

1,304.7
Mean
1,350
Median
160.8
Std dev
1,057.9
10th percentile
1,467.2
90th percentile
740.3
Min
740.31,569.6
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.21 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 data1,569.6
2Gemini 4 ArgonHigh· LMArena data1,544.7
3Gemini 4 NeonHigh· LMArena data1,525.3
4Claude Opus 4.6High· LMArena data1,511.7
5Muse WonderMax· LMArena data1,511.6
6Claude Opus 4.7High· LMArena data1,505.8
7Claude Opus 4.6· LMArena data1,505.1
8Claude Opus 4.7· LMArena data1,504.2
9Claude Fable 5High· LMArena data1,504.1
10Gemini 3.8 FlashHigh· LMArena data1,502
11Claude Opus 5.5High· LMArena data1,502
12Claude Opus 5Max· LMArena data1,498.9
13Muse Spark 1.2xHigh· LMArena data1,495.4
14Claude Opus 5High· LMArena data1,493.6
15Qwen3.8 Max· LMArena data1,492.1
16Claude Fable 5.1Max· LMArena data1,489.6
17Gemini 4 HydrogenHigh· LMArena data1,488.9
18Kimi K3Max· LMArena data1,488.8
19Muse Spark 1.3Max· LMArena data1,488.1
20Gemini 3.7 FlashHigh· LMArena data1,487.4
21Gemini 3.1 Pro Preview· LMArena data1,485.5
22Muse Spark 1.1· LMArena data1,484.7
23Gemini 3 Pro Preview· LMArena data1,483.8
24Gemini 3.6 FlashHigh· LMArena data1,483.3
25GPT 5.4High· LMArena data1,482.1
26Gemini 3.5 FlashHigh· LMArena data1,480.7
27Qwen3.7 Max Preview· LMArena data1,478
28Qwen3.5-Max-Preview· LMArena data1,477.8
29Gemini 3.5 FlashMedium· LMArena data1,477.6
30GLM 5.3Max· LMArena data1,477.4
31Muse Spark· LMArena data1,477.1
32GPT 5.5High· LMArena data1,476.9
33Claude Opus 4.8High· LMArena data1,476.3
34MiMo V2.5 Pro· LMArena data1,475.4
35GLM 5.3 Flash· LMArena data1,474
36GPT 5.5· LMArena data1,472.9
37GLM 5.1· LMArena data1,472.5
38MiMo V2.6 Pro· LMArena data1,471.9
39Gemini 3 Flash· LMArena data1,471.6
40Claude Opus 4.8· LMArena data1,471.3
41ERNIE 5.1· LMArena data1,471.2
42GLM 5.2Max· LMArena data1,469.6
43DeepSeek V4 Pro· LMArena data1,467.2
44Claude Opus 4.520251101· LMArena data1,466.3
45Claude Sonnet 4.520250929· LMArena data1,464.7
46GPT 5.4· LMArena data1,463.3
47Claude Sonnet 4.6· LMArena data1,462.6
48Claude Sonnet 5.5xHigh· LMArena data1,462.1
49GPT-5.6 SolxHigh· LMArena data1,461.9
50DeepSeek V4.1 Flash 0910Max· LMArena data1,461.1

Showing 1–50 of 421 results.

Data credits: LMArena, CC BY 4.0 (licence)