Skip to content

Benchmark

Arena ELO Text Factuality - Hard Prompts English

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

Models measured
156
Scores recorded
175
Sources
1
IQ weight
2.08

Score distribution

1,472.5
Mean
1,475.6
Median
36.4
Std dev
1,421.8
10th percentile
1,513.7
90th percentile
1,363.1
Min
1,363.11,545.9
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.08 for chat models, measured over 156 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51GPT-5.6 LunaxHigh· LMArena data1,496.7
52GLM 5.2Max· LMArena data1,496.1
53MiMo V2 Pro· LMArena data1,495.6
54Gemini 3.1 Pro Preview· LMArena data1,495.5
55Claude Sonnet 4.520250929· LMArena data1,495.5
56Muse Spark· LMArena data1,494.3
57Muse Spark 1.1· LMArena data1,494.2
58GPT-6 LunaMax· LMArena data1,494
59Qwen3.7 Max Preview· LMArena data1,493
60ERNIE 5.1· LMArena data1,492.7
61Gemini 3.5 FlashMedium· LMArena data1,492.6
62GPT 5.1High· LMArena data1,491.9
63Beam 1.0 501B A23B· LMArena data1,489.9
64Kimi K2.6· LMArena data1,489.4
65DeepSeek V4 ProHigh, 20260813· LMArena data1,488.7
66DeepSeek V4 Pro· LMArena data1,488
67Claude Sonnet 5High· LMArena data1,486.7
68GLM 5· LMArena data1,485.2
69Kimi K2.5Instant· LMArena data1,484.8
70Step 5 Preview· LMArena data1,484.6
71Qwen3.7 Plus· LMArena data1,482.9
72Grok 4.20Reasoning, 0309, beta· LMArena data1,482.9
73Gemini 3 Flash· LMArena data1,482.7
74MiMo V2.5· LMArena data1,481.7
75DeepSeek-V4-Pro PreviewHigh· LMArena data1,481.6
76Claude Opus 4.1Thinking 16K, 20250805· LMArena data1,481.5
77Grok 4.20 Multi-Agent0309, beta· LMArena data1,481.1
78Grok 4.6High· LMArena data1,480.7
79MiMo V2 Omni· LMArena data1,480.7
80GPT-5.2 Chat20260210· LMArena data1,479.1
81Inkling FP4· LMArena data1,478.8
82Qwen 3.8 27B· LMArena data1,478.5
83Hy3 (Deep Infra)· LMArena data1,478.1
84GPT 5.1· LMArena data1,478
85Kimi K2.5Thinking· LMArena data1,477.3
86LongCat-Flash-Chat Exp2602· LMArena data1,477
87Gemini 2.5 Pro· LMArena data1,475.9
88DeepSeek V3.2· LMArena data1,475.6
89Qwen3.5 397B A17B· LMArena data1,474.5
90Gemma 4 26B A4B· LMArena data1,473.9
91Qwen3.6 Plus· LMArena data1,473.5
92Claude Opus 4.120250805· LMArena data1,473.4
93DeepSeek V3.2Thinking· LMArena data1,472.8
94MiniMax M3· LMArena data1,472.3
95Mistral Medium 3.5· LMArena data1,472.2
96Gemini 3.5 Flash Lite· LMArena data1,472
97DeepSeek V4 Flash· LMArena data1,471.5
98GPT-5.4 miniHigh· LMArena data1,470.9
99Grok 4.20 Beta 1· LMArena data1,470.7
100MiniMax M2.7· LMArena data1,469.7

Showing 51–100 of 175 results.

Data credits: LMArena, CC BY 4.0 (licence)