Skip to content

Benchmark

Arena ELO Text Factuality - Industry Legal and Government

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

Models measured
116
Scores recorded
132
Sources
1
IQ weight
2.45

Score distribution

1,464.6
Mean
1,463.8
Median
32.6
Std dev
1,428
10th percentile
1,507
90th percentile
1,353.2
Min
1,353.21,526.1
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.45 for chat models, measured over 116 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51DeepSeek V4.1 Flash 0910Max· LMArena data1,474.6
52Gemini 3 Flash· LMArena data1,474.6
53Claude Sonnet 5High· LMArena data1,472.7
54GPT-6 LunaMax· LMArena data1,471.9
55Kimi K2.6· LMArena data1,471.6
56Gemini 3 FlashMinimal· LMArena data1,470.4
57GLM 4.6· LMArena data1,469.4
58DeepSeek-V4-Pro PreviewHigh· LMArena data1,469.4
59GPT-5.2 Chat20260210· LMArena data1,468.6
60Inkling FP4· LMArena data1,467.9
61GPT 5.2High· LMArena data1,467.6
62Grok 4.20Reasoning, 0309, beta· LMArena data1,467.1
63ERNIE 5.1· LMArena data1,466.6
64Claude Sonnet 4.520250929· LMArena data1,466.5
65GLM 5· LMArena data1,466.1
66Grok 4.20 Multi-Agent0309, beta· LMArena data1,463.8
67Grok 4.6High· LMArena data1,462.2
68Qwen3.5 397B A17B· LMArena data1,462.2
69GPT-5.4 miniHigh· LMArena data1,461.5
70Step 5 Preview· LMArena data1,461.3
71GPT-6 SolMax· LMArena data1,460.9
72GPT 5.2· LMArena data1,460.5
73MiMo V2.6 Flash· LMArena data1,458.8
74Gemini 2.5 Flash Preview09-2025· LMArena data1,458.4
75ChatGPT-4o Latest20250326· LMArena data1,458.2
76GPT-5.5 Instant· LMArena data1,457.5
77Claude Sonnet 4.5High, 32K, 20250929· LMArena data1,456.7
78Qwen3.7 Plus· LMArena data1,456.4
79MiMo V2 Pro· LMArena data1,456.3
80Grok 4.7xHigh· LMArena data1,455.9
81Gemini 3.5 Flash Lite· LMArena data1,454.5
82DeepSeek V4 Flash· LMArena data1,453.6
83DeepSeek V3.2Thinking· LMArena data1,453.3
84Qwen3 235B A22B Instruct 25072507· LMArena data1,453.2
85Gemini 2.5 Flash· LMArena data1,453.2
86Gemini 3.1 Flash Lite· LMArena data1,452.3
87ERNIE 5.00110· LMArena data1,450.7
88DeepSeek-V4-Flash PreviewHigh· LMArena data1,450.6
89Qwen 3.8 27B· LMArena data1,450.5
90Grok 4.1· LMArena data1,450.1
91MiMo V2 Omni· LMArena data1,449.6
92Hy3 (Deep Infra)· LMArena data1,449.5
93Mistral Large 3· LMArena data1,449
94Kimi K2.5Thinking· LMArena data1,448.2
95Mistral Medium 3.12508· LMArena data1,447.9
96Qwen3.6 Plus· LMArena data1,446.4
97Claude Opus 4.1Thinking 16K, 20250805· LMArena data1,445.8
98Grok 4.20 Beta 1· LMArena data1,445.8
99Claude Opus 4.120250805· LMArena data1,444.4
100DeepSeek V3.2 Exp· LMArena data1,443.9

Showing 51–100 of 132 results.

Data credits: LMArena, CC BY 4.0 (licence)