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
51Claude Opus 4.5High, 32K, 20251101· LMArena data1,460.2
52Qwen3.7 Plus· LMArena data1,459.7
53GPT-5.2 Chat20260210· LMArena data1,458.2
54Claude Fable 5.2Max· LMArena data1,457.8
55MiMo V2 Pro· LMArena data1,456.5
56GLM 5· LMArena data1,455.7
57Grok 4.20Reasoning, 0309, beta· LMArena data1,455.6
58Gemini 3 FlashMinimal· LMArena data1,455.3
59Grok 4.5· LMArena data1,454.8
60GPT-5.6 TerraxHigh· LMArena data1,454.7
61ChatGPT-4o Latest20250326· LMArena data1,454.3
62Qwen3.6 Max Preview· LMArena data1,454
63Kimi K2.6· LMArena data1,453.9
64Claude Sonnet 5High· LMArena data1,453.8
65Claude Sonnet 4.5High, 32K, 20250929· LMArena data1,453.1
66DeepSeek V4 ProHigh, 20260813· LMArena data1,452.4
67Gemma 4 31B IT· LMArena data1,452.3
68Gemini 2.5 Pro· LMArena data1,452.3
69Grok 4.20 Multi-Agent0309, beta· LMArena data1,451.9
70Grok 4.20 Beta 1· LMArena data1,450.3
71GPT-6.1 SolMax· LMArena data1,449.8
72GPT 5.1High· LMArena data1,449.8
73Kimi K3.1 ProMax· LMArena data1,448.8
74DeepSeek V4 Flash· LMArena data1,448.6
75GLM 4.7· LMArena data1,446.2
76Qwen3 Max Preview· LMArena data1,446.1
77Gemini 3.5 Flash Lite· LMArena data1,445.9
78GPT-6 AstraMax· LMArena data1,445.8
79Qwen3.5 397B A17B· LMArena data1,445.4
80DeepSeek-V4-Pro PreviewHigh· LMArena data1,445.2
81MiMo V2 Omni· LMArena data1,445.1
82MiMo V2.5· LMArena data1,444.5
83MiMo V2.6 Flash· LMArena data1,444.3
84Kimi K2.5Thinking· LMArena data1,444.1
85GPT-4.5 Preview2025-02-27· LMArena data1,444
86Claude Opus 4.1Thinking 16K, 20250805· LMArena data1,444
87Claude Opus 4.120250805· LMArena data1,443
88nova-experimental-chat26-02-10· LMArena data1,442.1
89Gemma 4 26B A4B· LMArena data1,441.2
90Doubao Seed 2.0 Pro· LMArena data1,441.2
91Inkling FP4· LMArena data1,440.4
92Step 5 Preview· LMArena data1,439.4
93Qwen 3.8 27B· LMArena data1,438.6
94Qwen3.6 Plus· LMArena data1,438.5
95Grok 4.1· LMArena data1,437.8
96Hy3 (Deep Infra)· LMArena data1,437.7
97MiniMax M3· LMArena data1,437.3
98GPT-5.6 LunaxHigh· LMArena data1,434.8
99GPT-5.5 Instant· LMArena data1,433.7
100Qwen3 Max2025-09-23· LMArena data1,433.3

Showing 51–100 of 421 results.

Data credits: LMArena, CC BY 4.0 (licence)