Skip to content

Benchmark

Arena ELO Text Factuality - Russian

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

Models measured
130
Scores recorded
146
Sources
1
IQ weight
2.22

Score distribution

1,450.9
Mean
1,456.3
Median
39.2
Std dev
1,407.8
10th percentile
1,497.3
90th percentile
1,300.5
Min
1,300.51,534.6
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.22 for chat models, measured over 130 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-5.2 Chat20260210· LMArena data1,465
52Grok 4.20 Multi-Agent0309, beta· LMArena data1,463.9
53ERNIE 5.1· LMArena data1,463.7
54Claude Opus 4.520251101· LMArena data1,463
55MiMo V2.5 Pro· LMArena data1,462.6
56GLM 5.1· LMArena data1,461.7
57DeepSeek V4 Pro· LMArena data1,461.3
58Grok 4.20 Beta 1· LMArena data1,461.1
59GPT-6 LunaMax· LMArena data1,461
60GPT 5.1High· LMArena data1,461
61DeepSeek V4 ProHigh, 20260813· LMArena data1,460.8
62GPT-5.6 LunaxHigh· LMArena data1,459.6
63Claude Sonnet 4.520250929· LMArena data1,459.4
64Kimi K2.6· LMArena data1,459.3
65Claude Sonnet 5High· LMArena data1,458.5
66GPT-5.4 miniHigh· LMArena data1,458
67ChatGPT-4o Latest20250326· LMArena data1,456.9
68Hy3 (Deep Infra)· LMArena data1,456.7
69Gemini 3.5 Flash Lite· LMArena data1,456.5
70GPT-6.1 TerraMax· LMArena data1,456.4
71Claude Opus 4.5High, 32K, 20251101· LMArena data1,456.4
72ERNIE 5.00110· LMArena data1,456.3
73DeepSeek-V4-Pro PreviewHigh· LMArena data1,456.3
74Inkling FP4· LMArena data1,456.1
75GLM 5· LMArena data1,454.7
76GPT-5.5 Instant· LMArena data1,454.1
77Claude Sonnet 4.6· LMArena data1,453.7
78Claude Sonnet 4.5High, 32K, 20250929· LMArena data1,452.7
79Qwen3.5 397B A17B· LMArena data1,450.8
80GPT-6 SolMax· LMArena data1,450.2
81GPT 5.2· LMArena data1,449.8
82GPT 5.1· LMArena data1,448.9
83Kimi K2.5Thinking· LMArena data1,448.5
84Claude Opus 4.120250805· LMArena data1,445.8
85Grok 4.6High· LMArena data1,445.7
86Step 5 Preview· LMArena data1,443.4
87Gemini 3.1 Flash Lite· LMArena data1,443.4
88Qwen3.6 Plus· LMArena data1,442.7
89DeepSeek V4 Flash· LMArena data1,442.7
90Grok 4.7xHigh· LMArena data1,441.7
91Grok 4.1Thinking· LMArena data1,441.4
92Doubao Seed 2.0 Pro· LMArena data1,441.4
93Inkling 2 Small· LMArena data1,441.2
94MiMo V2.6 Flash· LMArena data1,440.7
95GPT 5.2High· LMArena data1,439.9
96MiMo V2 Pro· LMArena data1,439.9
97DeepSeek-V4-Flash PreviewHigh· LMArena data1,439.5
98Claude Opus 4.1Thinking 16K, 20250805· LMArena data1,437.9
99DeepSeek V3.2· LMArena data1,437.2
100GLM 4.6· LMArena data1,436.5

Showing 51–100 of 146 results.

Data credits: LMArena, CC BY 4.0 (licence)