Skip to content

Benchmark

Arena ELO Text Factuality - Expert

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

Models measured
114
Scores recorded
130
Sources
1
IQ weight
2.21

Score distribution

1,480.4
Mean
1,482
Median
33.7
Std dev
1,437.1
10th percentile
1,522
90th percentile
1,386.2
Min
1,386.21,552.1
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.21 for chat models, measured over 114 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51Gemini 3.1 Pro Preview· LMArena data1,492.8
52MiMo V2 Pro· LMArena data1,492.8
53Grok 4.6High· LMArena data1,490.2
54Claude Sonnet 4.520250929· LMArena data1,488.8
55Inkling FP4· LMArena data1,487.3
56DeepSeek V4 Pro· LMArena data1,486.8
57GLM 5.1· LMArena data1,486.6
58GPT-6 SolMax· LMArena data1,486.5
59Qwen3.5 397B A17B· LMArena data1,485.4
60GPT 5.1High· LMArena data1,485.4
61Gemini 3 Flash· LMArena data1,484.5
62Qwen3.7 Plus· LMArena data1,484.1
63Qwen 3.8 27B· LMArena data1,483.3
64Muse Spark 1.1· LMArena data1,483.2
65Hy3 (Deep Infra)· LMArena data1,482
66Kimi K2.5Thinking· LMArena data1,481.3
67GPT 5.1· LMArena data1,480.8
68DeepSeek-V4-Pro PreviewHigh· LMArena data1,479.4
69MiMo V2.5· LMArena data1,478.1
70GLM 5· LMArena data1,477
71Grok 4.20 Multi-Agent0309, beta· LMArena data1,476.7
72GPT 5.2High· LMArena data1,476.3
73Nemotron 3 UltraNVFP4· LMArena data1,475.9
74MiniMax M3· LMArena data1,475.1
75DeepSeek V4 ProHigh, 20260813· LMArena data1,474.4
76Qwen3 235B A22B Instruct 25072507· LMArena data1,474
77Gemini 2.5 Pro· LMArena data1,473.9
78GPT-5.4 miniHigh· LMArena data1,473.2
79Mistral Large 4· LMArena data1,471.6
80MiMo V2 Omni· LMArena data1,470.6
81Claude Haiku 5.5High· LMArena data1,468.7
82LongCat-Flash-Chat Exp2602· LMArena data1,468
83Qwen3.6 Plus· LMArena data1,467.9
84GPT 5.2· LMArena data1,466
85Claude Haiku 4.520251001· LMArena data1,464.9
86GPT-5.2 Chat20260210· LMArena data1,464.8
87MiniMax M2.7· LMArena data1,463.9
88Gemini 3.5 Flash Lite· LMArena data1,462.7
89Claude Opus 4.120250805· LMArena data1,462.3
90GLM 4.6· LMArena data1,462.1
91Grok 4.20Reasoning, 0309, beta· LMArena data1,461.9
92DeepSeek V4 Flash· LMArena data1,461.9
93Grok 4.7xHigh· LMArena data1,461.7
94Solar Pro 4· LMArena data1,460
95Inkling Small· LMArena data1,459.7
96DeepSeek-V4-Flash PreviewHigh· LMArena data1,459.5
97DeepSeek V3.2· LMArena data1,456.8
98MiniMax-M2.1 Preview· LMArena data1,456.6
99Grok 4.20 Beta 1· LMArena data1,455.8
100Gemini 2.5 Flash· LMArena data1,455.7

Showing 51–100 of 130 results.

Data credits: LMArena, CC BY 4.0 (licence)