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
101GLM 4.5· LMArena data1,379
102Qwen3 Max2025-09-23· LMArena data1,378.7
103Qwen3.6 Plus· LMArena data1,377.7
104Grok 40709· LMArena data1,376
105Claude Haiku 5.5High· LMArena data1,375.3
106GPT 5.2· LMArena data1,375
107Qwen 3 235B A22B2507, thinking· LMArena data1,374.7
108Grok 3 Preview02-24· LMArena data1,373.2
109Claude Opus 4.5High, 32K, 20251101· LMArena data1,373
110ERNIE-5.0-Preview1203· LMArena data1,371.6
111DeepSeek V3.2 Exp· LMArena data1,370.7
112DeepSeek V3.2· LMArena data1,370.3
113OpenAI o32025-04-16· LMArena data1,369.7
114GPT 5.2High· LMArena data1,368.6
115LongCat-Flash-Chat Exp2602· LMArena data1,368.5
116Grok 4.7xHigh· LMArena data1,368.2
117GPT-5.4 miniHigh· LMArena data1,367.8
118Claude Sonnet 4.5High, 32K, 20250929· LMArena data1,367.6
119Claude Opus 4.120250805· LMArena data1,363.7
120Qwen3 Next 80B A3B Instruct· LMArena data1,363.2
121DeepSeek V3.2Thinking· LMArena data1,363.1
122MiniMax M3· LMArena data1,362.1
123Grok 4.1 FastReasoning· LMArena data1,360.6
124Gemini 2.5 Flash-Lite PreviewThinking, 06-17· LMArena data1,360.5
125GPT-5 Chat· LMArena data1,359.7
126DeepSeek R10528· LMArena data1,358.9
127DeepSeek-V4-Flash PreviewHigh· LMArena data1,358.8
128MiMo V2 FlashNon-Thinking· LMArena data1,358.7
129GPT 5High· LMArena data1,358.6
130Qwen 3 235B A22BNon-Thinking· LMArena data1,358.5
131Grok 4 FastReasoning· LMArena data1,358.2
132LongCat Flash Chat· LMArena data1,358
133Qwen3.5 27B· LMArena data1,356.4
134MiMo V2.5· LMArena data1,356.1
135Qwen3.5 35B A3B· LMArena data1,354.7
136Qwen3.5 122B-A10B· LMArena data1,354.7
137Qwen3 VL 235Bthinking· LMArena data1,354.6
138Grok 4.3· LMArena data1,352.4
139MiMo V2 Omni· LMArena data1,352.2
140Gemini 2.5 Flash-Lite PreviewNon-Thinking, 09-2025· LMArena data1,351.9
141Hunyuan TurboS20250416· LMArena data1,351.3
142Step 3.5 Flash· LMArena data1,349.8
143DeepSeek V3.2 ExpThinking· LMArena data1,348.8
144GPT-6 LunaMax· LMArena data1,348.8
145GPT-5.3 Chat· LMArena data1,348.5
146Qwen3.5 Flash· LMArena data1,345.5
147Claude Haiku 4.520251001· LMArena data1,344.6
148GPT-6 SolMax· LMArena data1,342.4
149GPT 4.12025-04-14· LMArena data1,339.4
150DeepSeek V3.1· LMArena data1,337.6

Showing 101–150 of 288 results.

Data credits: LMArena, CC BY 4.0 (licence)