Skip to content

Benchmark

Arena ELO Text Factuality - Industry Mathematical

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

Models measured
102
Scores recorded
117
Sources
1
IQ weight
2.25

Score distribution

1,453.4
Mean
1,452.4
Median
30.2
Std dev
1,416.5
10th percentile
1,489.5
90th percentile
1,363.5
Min
1,363.51,520.2
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.25 for chat models, measured over 102 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51GLM 5.1· LMArena data1,459.1
52GLM 5· LMArena data1,458.1
53DeepSeek-V4-Pro PreviewHigh· LMArena data1,457.4
54GPT-6 LunaMax· LMArena data1,455.7
55GPT 5.2High· LMArena data1,454.2
56Qwen3.5 397B A17B· LMArena data1,453.1
57Grok 4.7xHigh· LMArena data1,452.7
58MiMo V2.5· LMArena data1,452.6
59MiMo V2 Omni· LMArena data1,452.4
60Gemini 3 Flash· LMArena data1,450.4
61GPT 5.1High· LMArena data1,448.7
62GLM 4.6· LMArena data1,448.4
63Qwen 3.8 27B· LMArena data1,448
64Claude Haiku 5.5High· LMArena data1,447.3
65Qwen3.6 Plus· LMArena data1,447.2
66DeepSeek V4 Pro· LMArena data1,447.2
67Qwen3 235B A22B Instruct 25072507· LMArena data1,446.8
68Claude Sonnet 4.520250929· LMArena data1,446.6
69Gemini 2.5 Pro· LMArena data1,446.2
70Grok 4.20 Multi-Agent0309, beta· LMArena data1,445.8
71Inkling Small· LMArena data1,445.6
72GPT-6 SolMax· LMArena data1,445.1
73Grok 4.20Reasoning, 0309, beta· LMArena data1,444.6
74MiniMax M3· LMArena data1,443.8
75GPT-5.4 miniHigh· LMArena data1,442.3
76LongCat-Flash-Chat Exp2602· LMArena data1,438
77Gemini 3 FlashMinimal· LMArena data1,437.9
78GPT 5.1· LMArena data1,437.9
79DeepSeek V3.2Thinking· LMArena data1,437.7
80Qwen3.5 27B· LMArena data1,436.6
81DeepSeek V4 Flash· LMArena data1,436.2
82MiniMax M2.7· LMArena data1,436.1
83ERNIE 5.00110· LMArena data1,436
84Claude Opus 4.1Thinking 16K, 20250805· LMArena data1,435.9
85Kimi K2 Thinking· LMArena data1,435.5
86Claude Haiku 4.520251001· LMArena data1,434.5
87Qwen3.5 122B-A10B· LMArena data1,434.2
88DeepSeek-V4-Flash PreviewHigh· LMArena data1,432.6
89DeepSeek V3.2· LMArena data1,432.6
90Solar Pro 4· LMArena data1,432.5
91Gemini 3.5 Flash Lite· LMArena data1,432.2
92GPT 5.2· LMArena data1,431
93GPT-5.2 Chat20260210· LMArena data1,431
94GPT 5.4 NanoHigh· LMArena data1,430.1
95Mistral Large 3· LMArena data1,429.2
96Step 3.5 Flash· LMArena data1,428.6
97Gemini 3.1 Flash Lite· LMArena data1,426.2
98Gemini 2.5 Flash Preview09-2025· LMArena data1,426
99Claude Opus 4.120250805· LMArena data1,425.9
100Mistral Medium 3.12508· LMArena data1,421.7

Showing 51–100 of 117 results.

Data credits: LMArena, CC BY 4.0 (licence)