LLM Stats (AlignBench): leaderboard
LLM Stats mirror of AlignBench, testing Chinese-language alignment across language understanding, writing, reasoning, role play, and professional knowledge.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Estimated already saturated. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 14B | 85.6 | #542 |
| 2 | Qwen 2.5 72B Instruct | 81.6 | #469 |
| 3 | DeepSeek V2.5 | 80.4 | #494 |
| 4 | Qwen 2.5 7B Instruct | 73.3 | #863 |
| 5 | Qwen 2 7B Instruct | 72.1 | #982 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-alignbench · How It Works · Data refreshed daily, snapshot 2026-11-05.