LiveBench Language category score (0–100). Evaluates sentence completion, language understanding, stylistic analysis, and translation quality.
Source: LiveBench| Rank | Model | |
|---|---|---|
| #1 | Gemini 3.1 Pro | 85.4 |
| #2 | Anthropic Claude Opus 4.6 | 83.3 |
| #3 | ChatGPT GPT-5.4 | 82.6 |
| #4 | Gemini 3 Flash | 78.7 |
| #5 | Moonshot AI Kimi K2.5 | 77.7 |
| #6 | Z.ai GLM-5 | 77.5 |
| #7 | Anthropic Claude Opus 4.5 | 77.2 |
| #8 | Anthropic Claude Sonnet 4.5 | 76.0 |
| #9 | Gemini 2.5 Pro | 75.5 |
| #10 | Qwen Qwen3.6 Plus | 75.0 |
| #11 | Gemini 3.1 Flash Lite | 73.2 |
| #12 | Z.ai GLM-5.1 | 71.8 |
| #13 | Anthropic Claude Sonnet 4.6 | 71.4 |
| #14 | Gemma 4 31B | 71.3 |
| #15 | ChatGPT GPT-5 Mini | 69.2 |
| #16 | Xiaomi MiMo-V2-Pro | 69.1 |
| #17 | MiniMax MiniMax M2.7 | 66.8 |
| #18 | DeepSeek DeepSeek V3.2 | 64.2 |
| #19 | Gemini 2.5 Flash | 62.3 |
| #20 | Anthropic Claude Haiku 4.5 | 57.0 |
| #21 | MiniMax MiniMax M2.5 | 55.1 |
| #22 | Gemini 2.5 Flash Lite | 52.0 |
| #23 | ChatGPT GPT OSS 120B | 48.6 |
| #24 | ChatGPT GPT-5 Nano | 47.7 |
| #25 | ChatGPT GPT-5.4 Mini | 41.8 |
| #26 | ChatGPT GPT-5.4 Nano | 28.7 |