Arena Elo score aggregated from user preference votes on Japanese-language prompts.
Source: Arena Intelligence| Rank | Model | |
|---|---|---|
| #1 | Gemini 3.1 Pro | 1533 |
| #2 | Anthropic Claude Opus 4.6 | 1512 |
| #3 | Gemini 3 Flash | 1486 |
| #4 | Qwen Qwen3.5 397B A17B | 1457 |
| #5 | Gemini 2.5 Pro | 1454 |
| #6 | Anthropic Claude Opus 4.5 | 1450 |
| #7 | Moonshot AI Kimi K2.5 | 1436 |
| #8 | ChatGPT GPT-5 | 1432 |
| #9 | Anthropic Claude Opus 4.1 | 1409 |
| #10 | Gemini 2.5 Flash | 1392 |
| #11 | Anthropic Claude Sonnet 4.5 | 1390 |
| #12 | Anthropic Claude Opus 4 | 1386 |
| #13 | Baidu ERNIE 5.0 Thinking | 1378 |
| #14 | ChatGPT GPT-5 Mini | 1369 |
| #15 | DeepSeek DeepSeek V3.2 | 1365 |
| #16 | Grok Grok 4.1 Fast | 1364 |
| #17 | Grok Grok 4.1 Fast (Reasoning) | 1364 |
| #18 | Gemini 2.5 Flash Lite | 1356 |
| #19 | Anthropic Claude Sonnet 4 | 1344 |
| #20 | Anthropic Claude Haiku 4.5 | 1338 |
| #21 | ChatGPT GPT OSS 120B | 1316 |
| #22 | ChatGPT GPT-4.1 | 1262 |
| #23 | Meta AI Llama 4 Scout | 1258 |
| #24 | Meta AI Llama 4 Maverick | 1253 |
| #25 | MiniMax MiniMax M2.5 | 1231 |