基於 SFC/HKMA 監管標準,對主流大語言模型進行五維度量化評測,數據每日更新並同步上鏈存證。
| # | MODEL | ETHICAL | TOTAL | TREND | SBT |
|---|---|---|---|---|---|
1 | Qwen-Max Alibaba | 100 | 93 | +3.2 | 待認證 |
2 | Qwen-Plus Alibaba | 100 | 93 | — | 待認證 |
3 | Qwen-Turbo Alibaba | 100 | 93 | — | 待認證 |
4 | Claude 3.5 Sonnet Anthropic | 94 | 90 | +2.1 | SBT ✓ |
5 | GPT-4o OpenAI | 88 | 88 | — | SBT ✓ |
6 | Qwen-Max (via OpenRouter) OpenRouter | 75 | 85 | — | 待認證 |
7 | Gemini 1.5 Pro Google | 86 | 84 | +1.3 | SBT ✓ |
8 | DeepSeek-V3 DeepSeek | 74 | 80 | -1.8 | 待認證 |
9 | Llama 3.3 70B Meta | 76 | 72 | +0.8 | 待認證 |
10 | Mistral Large Mistral AI | 72 | 68 | — | 待認證 |