Skip to content

Benchmark

Arena ELO - Korean

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

Models measured
244
Scores recorded
288
Sources
1
IQ weight
1.99

Score distribution

1,298.9
Mean
1,348.8
Median
139.4
Std dev
1,087.1
10th percentile
1,441.2
90th percentile
856
Min
8561,509.5
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.99 for chat models, measured over 244 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51Grok 4.20 Beta 1· LMArena data1,418.9
52Grok 4.20 Multi-Agent0309, beta· LMArena data1,416
53Qwen3.7 Plus· LMArena data1,415.6
54GPT-6 AstraMax· LMArena data1,415.1
55Claude Sonnet 4.6· LMArena data1,414
56MiMo V2.6 Flash· LMArena data1,412.5
57Inkling FP4· LMArena data1,412.2
58Kimi K2.5Thinking· LMArena data1,410.3
59Grok 4.1· LMArena data1,409.6
60GPT-5.6 LunaxHigh· LMArena data1,409.1
61ERNIE 5.00110· LMArena data1,407.6
62DeepSeek V4 Pro· LMArena data1,407.3
63Grok 4.5· LMArena data1,406.1
64Claude Opus 4.8· LMArena data1,405.9
65Hunyuan T120250711· LMArena data1,405.8
66Gemini 3.5 Flash Lite· LMArena data1,405.4
67Step 5 Preview· LMArena data1,402.5
68Hy3 (Deep Infra)· LMArena data1,401.8
69MiMo V2 Pro· LMArena data1,400.9
70GLM 4.6· LMArena data1,400.5
71Claude Sonnet 4.520250929· LMArena data1,400.3
72Claude Sonnet 5High· LMArena data1,399.9
73GPT 5.1High· LMArena data1,399.5
74Qwen 3.8 27B· LMArena data1,398.4
75GLM 4.7· LMArena data1,397.6
76Qwen3 Max Preview· LMArena data1,396.7
77Grok 4.1Thinking· LMArena data1,396.2
78o12024-12-17· LMArena data1,396.1
79ChatGPT-4o Latest20250326· LMArena data1,395.5
80GPT-5.5 Instant· LMArena data1,393.9
81Qwen3 VL 235B· LMArena data1,392.3
82GPT-5.2 Chat20260210· LMArena data1,392.3
83GPT-4.5 Preview2025-02-27· LMArena data1,392
84Grok 4.6High· LMArena data1,392
85Gemini 3.1 Flash Lite· LMArena data1,391.9
86Qwen3.5 397B A17B· LMArena data1,391.2
87Mistral Large 3· LMArena data1,391
88Qwen3 235B A22B Instruct 25072507· LMArena data1,390
89GPT 5.1· LMArena data1,389.6
90Gemini 2.5 Flash· LMArena data1,386.3
91Mistral Medium 3.5· LMArena data1,385.8
92nova-experimental-chat11-10· LMArena data1,384.5
93Gemini 2.5 Flash Preview09-2025· LMArena data1,384.4
94Inkling Small· LMArena data1,384
95Nemotron 3 UltraNVFP4· LMArena data1,383.4
96DeepSeek V4 Flash· LMArena data1,382.7
97Kimi K2.5Instant· LMArena data1,381.3
98Claude Opus 4.1Thinking 16K, 20250805· LMArena data1,381.2
99Solar Pro 4· LMArena data1,380.9
100Mistral Medium 3.12508· LMArena data1,380.3

Showing 51–100 of 288 results.

Data credits: LMArena, CC BY 4.0 (licence)