模型能力榜
来自 Artificial Analysis 公开榜快照 · 越高越好
展示前 27 名
点击柱或标签打开 Artificial Analysis 来源页
- Anthropic
- OpenAI
- Meta
- Moonshot
- xAI
- Tencent
- Alibaba
- DeepSeek
- Zhipu
- Thinking Machines
- Xiaomi
- InclusionAI
- Mistral
| 名次 | 名次 Δ | 模型 | 开发商 | 分数 | 关注 |
|---|---|---|---|---|---|
| 1 | — | Claude Fable 5.2 (max) | Anthropic | 64.6 | |
| 2 | ↓1 | Gemini 4 Carbon (high) | 60.0 | ||
| 3 | ↓1 | Claude Opus 5.5 (max) | Anthropic | 59.1 | |
| 4 | ↓1 | GPT-6 Astra (max) | OpenAI | 56.9 | |
| 5 | ↓1 | GPT-6.1 Sol (max) | OpenAI | 54.6 | |
| 6 | ↓1 | Claude Sonnet 5.5 (max) | Anthropic | 54.5 | |
| 7 | ↓1 | GPT-6.1 Terra (max) | OpenAI | 51.3 | |
| 8 | ↓1 | Muse Wonder Full (max) | Meta | 50.1 | |
| 9 | — | Kimi K3.1 Pro | Moonshot | 46.5 | |
| 10 | — | Muse Wonder (max) | Meta | 43.6 | |
| 11 | — | Claude Haiku 5.5 (max) | Anthropic | 36.5 | |
| 12 | — | Muse Ember (max) | Meta | 36.0 | |
| 13 | ↓1 | Grok 4.7 (xhigh) | xAI | 35.5 | |
| 14 | ↓1 | Tencent Hy4 | Tencent | 34.6 | |
| 15 | ↓1 | Qwen3.8 Max (0902) | Alibaba | 33.7 | |
| 16 | ↓1 | DeepSeek V4.1 Flash 1017 (max) | DeepSeek | 32.6 | |
| 17 | ↓1 | GLM-5.3 (max) | Zhipu | 31.7 | |
| 18 | ↓1 | Gemini 3.8 Flash (high) | 30.8 | ||
| 19 | ↓1 | Inkling 2 Small | Thinking Machines | 30.6 | |
| 20 | ↓1 | MiMo-V2.6-Pro | Xiaomi | 29.9 | |
| 21 | ↓1 | Ling 3.1 Flash 1019 | InclusionAI | 29.6 | |
| 22 | — | GPT-6 Luna (max) | OpenAI | 27.9 | |
| 23 | — | Mistral Large 4 (high) | Mistral | 26.1 | |
| 24 | ↑1 | Qwen3.8 27B (xhigh) | Alibaba | 20.1 | |
| 25 | ↑1 | K2 Horizon 375B A23B | Institute of Foundation Models | 17.8 | |
| 26 | ↑1 | Beam 1.0 | Reflection | 17.7 | |
| 27 | ↑1 | Nemotron 3.5 Super VL | NVIDIA | 13.6 |
