Skip to content

Benchmark

Arena ELO Text Factuality - Chinese

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

Models measured
108
Scores recorded
123
Sources
1
IQ weight
2.06

Score distribution

1,492.7
Mean
1,492
Median
35.3
Std dev
1,451.2
10th percentile
1,534.8
90th percentile
1,389.6
Min
1,389.61,562.4
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.06 for chat models, measured over 108 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51Inkling FP4· LMArena data1,502
52DeepSeek-V4-Pro PreviewHigh· LMArena data1,501
53Qwen3.5 397B A17B· LMArena data1,500.5
54Kimi K2.5Thinking· LMArena data1,500.2
55Hy3 (Deep Infra)· LMArena data1,499.5
56GPT 5.1· LMArena data1,499
57GPT-6 LunaMax· LMArena data1,498.6
58Gemini 2.5 Pro· LMArena data1,498.3
59GPT-5.6 LunaxHigh· LMArena data1,497.6
60Claude Sonnet 4.6· LMArena data1,496.8
61ERNIE 5.00110· LMArena data1,495.7
62DeepSeek V4 Pro· LMArena data1,492
63Qwen 3.8 27B· LMArena data1,491.1
64Grok 4.6High· LMArena data1,490.1
65Grok 4.20 Multi-Agent0309, beta· LMArena data1,489.3
66Grok 4.20Reasoning, 0309, beta· LMArena data1,487.7
67Claude Opus 4.5High, 32K, 20251101· LMArena data1,487.6
68Qwen3 235B A22B Instruct 25072507· LMArena data1,487.2
69Gemini 3.5 Flash Lite· LMArena data1,486.4
70Claude Sonnet 5High· LMArena data1,486.4
71Gemini 3 FlashMinimal· LMArena data1,485.1
72MiMo V2.5· LMArena data1,484.4
73Claude Opus 4.520251101· LMArena data1,484.2
74Gemini 2.5 Flash Preview09-2025· LMArena data1,483.6
75Qwen3.5 27B· LMArena data1,483.2
76Claude Sonnet 4.520250929· LMArena data1,482.8
77GPT-5.5 Instant· LMArena data1,482.3
78Claude Sonnet 4.5High, 32K, 20250929· LMArena data1,479
79MiniMax M3· LMArena data1,478.6
80DeepSeek-V4-Flash PreviewHigh· LMArena data1,477.3
81GPT 5.2High· LMArena data1,477.2
82Qwen3.5 122B-A10B· LMArena data1,476.1
83LongCat-Flash-Chat Exp2602· LMArena data1,475
84DeepSeek V4 Flash· LMArena data1,474.8
85Doubao Seed 2.0 Pro· LMArena data1,474.4
86Qwen3.6 Plus· LMArena data1,473.5
87Gemini 3.1 Flash Lite· LMArena data1,473.4
88GPT-5.4 miniHigh· LMArena data1,472.1
89GPT 5.2· LMArena data1,471.3
90MiMo V2 FlashNon-Thinking· LMArena data1,471.3
91DeepSeek V3.2· LMArena data1,471.1
92Inkling Small· LMArena data1,470.8
93Grok 4.7xHigh· LMArena data1,470.6
94Qwen3.5 35B A3B· LMArena data1,469.5
95Kimi K2 Thinking· LMArena data1,469
96Claude Haiku 5.5High· LMArena data1,469
97Gemini 2.5 Flash· LMArena data1,467.5
98DeepSeek V3.2Thinking· LMArena data1,467.1
99GPT-5.2 Chat20260210· LMArena data1,466.4
100Grok 4.1· LMArena data1,465.1

Showing 51–100 of 123 results.

Data credits: LMArena, CC BY 4.0 (licence)