Skip to content

Benchmark

Arena ELO - Industry Legal and Government

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

Models measured
338
Scores recorded
394
Sources
1
IQ weight
1.34

Score distribution

1,329.7
Mean
1,368.1
Median
139.4
Std dev
1,104.4
10th percentile
1,475.2
90th percentile
895.4
Min
895.41,543.1
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.34 for chat models, measured over 338 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51Grok 4.5· LMArena data1,464.2
52DeepSeek V4 Pro· LMArena data1,463.8
53GPT 5.4· LMArena data1,462.3
54GPT-5.6 TerraxHigh· LMArena data1,462.3
55DeepSeek V4 ProHigh, 20260813· LMArena data1,462
56Kimi K2.6· LMArena data1,461.2
57GPT 5.1High· LMArena data1,459.4
58DeepSeek V3.1 TerminusThinking· LMArena data1,459.1
59Claude Opus 4.520251101· LMArena data1,459.1
60ERNIE-5.0-Preview1022· LMArena data1,458.7
61DeepSeek-V4-Pro PreviewHigh· LMArena data1,457.5
62Gemma 4 26B A4B· LMArena data1,456.8
63GPT-6.1 TerraMax· LMArena data1,454.5
64Nemotron 3 UltraNVFP4· LMArena data1,454.4
65Grok 4.20Reasoning, 0309, beta· LMArena data1,454.2
66GPT-6 AstraMax· LMArena data1,454.2
67Claude Opus 4.5High, 32K, 20251101· LMArena data1,453.7
68GLM 5· LMArena data1,453
69Qwen3.6 Max Preview· LMArena data1,451.9
70Qwen3.7 Plus· LMArena data1,451.9
71Claude Sonnet 5High· LMArena data1,450.8
72Inkling FP4· LMArena data1,449.7
73Claude Sonnet 4.520250929· LMArena data1,449.7
74Doubao Seed 2.0 Pro· LMArena data1,449.6
75Grok 4.20 Multi-Agent0309, beta· LMArena data1,448.4
76GLM 4.6· LMArena data1,448.2
77ERNIE-5.0-Preview1203· LMArena data1,448.2
78DeepSeek V4.1 Flash 0910Max· LMArena data1,447.1
79MiMo V2 Pro· LMArena data1,446.9
80Qwen 3.8 27B· LMArena data1,446.9
81MiMo V2.6 Flash· LMArena data1,445.7
82Gemini 3 FlashMinimal· LMArena data1,445.6
83Grok 4.1· LMArena data1,445.5
84Grok 3 Preview02-24· LMArena data1,444.9
85ChatGPT-4o Latest20250326· LMArena data1,444.8
86GLM 4.7· LMArena data1,444.7
87GPT-5.6 LunaxHigh· LMArena data1,444.5
88Qwen3.5 397B A17B· LMArena data1,442.4
89Claude Sonnet 4.5High, 32K, 20250929· LMArena data1,441.5
90Mistral Large 3· LMArena data1,441.3
91GPT 5.1· LMArena data1,440.8
92Kimi K2.5Thinking· LMArena data1,440.4
93Mistral Medium 3.12508· LMArena data1,439.9
94Step 5 Preview· LMArena data1,438.6
95Qwen3 VL 235B· LMArena data1,438.5
96DeepSeek-V4-Flash PreviewHigh· LMArena data1,438.5
97Grok 4.1Thinking· LMArena data1,438.3
98DeepSeek V4 Flash· LMArena data1,438.2
99Gemma 4 31B IT· LMArena data1,438.1
100GLM 4.5· LMArena data1,437.9

Showing 51–100 of 394 results.

Data credits: LMArena, CC BY 4.0 (licence)