| 排名 | 模型 | 上下文大小 | 开发商 | 智能指数 |
|---|---|---|---|---|
| #1 | DeepSeek V4 Pro (max) | 1M | 45 | |
| #2 | GLM-5.3 (max) | 1M | 45 | |
| #3 | Kimi K3 (max) | 1.05M | 44 | |
| #4 | Qwen3.8-Flash-Next | 256k | 42 | |
| #5 | GLM-5.3-Flash | 1M | 42 | |
| #6 | Qwen3.8 Max | 1M | 40 | |
| #7 | Qwen3.8 2.4T A95B | 984k | 40 | |
| #8 | GLM-5.2 (max) | 1M | 39 | |
| #9 | DeepSeek V4 Pro 0813 (Reasoning, Max Effort) | 1M | 36 | |
| #10 | DeepSeek V4 Flash Vision (Reasoning, Max Effort) | 1M | 35 | |
| #11 | DeepSeek V4 Flash 0731 (Reasoning, Max Effort) | 1M | 35 | |
| #12 | Kimi K3 (low) | 1.05M | 35 | |
| #13 | Qwen3.8 27B (xhigh) | 256k | 34 | |
| #14 | DeepSeek V4 Pro | 1M | 32* | |
| #15 | DeepSeek V4 Pro (Reasoning, Max Effort) | 1M | 31 | |
| #16 | Qwen3.8 27B (medium) | 256k | 31 | |
| #17 | DeepSeek V4 Pro (Reasoning, High Effort) | 1M | 30 | |
| #18 | MiniMax-M3 | 1M | 30 | |
| #19 | Qwen3.8 27B (low) | 256k | 29 | |
| #20 | DeepSeek V4 Flash | 1M | 29* |