| 模型 | 上下文 | 输入价 | 输出价 | 吞吐 | 延迟 | 服务商 | Agents | 编程 | 推理 | 知识 | 数学 | 多语言 | 多模态 | 指令遵循 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Granite 4.1 8B | 131.1K | — | — | — | — | — | — | 21.8 | 23.5 | — | — | — | — | — |
| Nemotron 3 Nano Omni | 256K | — | — | — | — | +8 | — | 27.8 | 26.1 | — | — | — | — | — |
| Qwen3.6 35B A3B | 262.1K | — | — | — | — | +1 | 55.9 | 50 | 50.7 | 51.3 | 87 | — | 80 | 64.4 |
| Qwen3.6 Max Preview | 262.1K | — | — | — | — | +9 | 80.1 | 50 | 47.8 | 46.2 | — | — | — | 76.6 |
| Qwen3.6 27B | 262.1K | — | — | — | — | 66.3 | 55 | 52.4 | 57.7 | 88.7 | — | 81.6 | 67.6 | |
| Hy3 preview | 262.1K | — | — | — | — | +1 | 61.3 | 47.8 | 42.9 | 58 | — | — | — | 63.1 |
| Nemotron 3 Super | 1M | — | — | — | — | +5 | — | — | — | — | — | — | — | — |
| Qwen3.5-9B | 262.1K | — | — | — | — | — | — | 27.7 | 43.6 | — | — | — | — | — |
| Gemini 3.1 Flash Lite Preview | 1.0M | — | — | — | — | — | 41.9 | 49.2 | — | — | — | — | — | |
| Qwen3.5-35B-A3B | 262.1K | — | — | — | — | — | 54.8 | 43.3 | 50.4 | 50.2 | — | 81 | 80.6 | 82.2 |
| Qwen3.5-27B | 262.1K | — | — | — | — | — | 58.4 | 48 | 52.1 | 51.3 | — | 82.2 | 82.1 | 85.3 |
| Qwen3.5-122B-A10B | 262.1K | — | — | — | — | 58.5 | 45.3 | 52 | 53.1 | — | 82.2 | 81.7 | 84.6 | |
| Gemini 3.1 Pro Preview | 1.0M | — | — | — | — | +2 | — | 58.9 | 69.4 | — | — | — | — | — |
| Qwen3.5 397B A17B | 256K | — | — | — | — | — | — | 41.1 | 52.5 | — | — | — | — | — |
| Gemini 3 Flash Preview | 1.0M | — | — | — | — | — | 70.7 | 71.2 | — | — | — | — | — | |
| Olmo 3 32B Think | 65.5K | — | — | — | — | — | — | 47.9 | 47.6 | — | — | — | — | — |
| Qwen3 VL 32B Instruct | 262.1K | — | — | — | — | — | 40.8 | 50.8 | — | — | — | — | — | |
| Qwen3 VL 8B Instruct | 256K | — | — | — | — | — | 25.3 | 38.1 | — | — | — | — | — | |
| Qwen3 VL 30B A3B Instruct | 262.1K | — | — | — | — | — | 39.2 | 50.8 | — | — | — | — | — | |
| DeepSeek V3.2 Exp | 163.8K | — | — | — | — | — | — | 47.7 | 55.3 | — | — | — | — | — |
| Gemini 2.5 Flash Lite Preview 09-2025 | 1.0M | — | — | — | — | — | — | 46.3 | 49.8 | — | — | — | — | — |
| Qwen3 VL 235B A22B Instruct | 262.1K | — | — | — | — | +1 | — | 47.7 | 53.3 | — | — | — | — | — |
| Qwen3 Next 80B A3B Instruct | 262.1K | — | — | — | — | +13 | — | 49.6 | 54.3 | — | — | — | — | — |
| Kimi K2 0905 | 262.1K | — | — | — | — | — | — | 45.9 | 55 | — | — | — | — | — |
| gpt-oss-120b | 131.1K | — | — | — | — | 37.7 | 45.2 | 40.3 | 46.6 | — | — | 100 | 69 | |
| gpt-oss-20b | 131.1K | — | — | — | — | — | 33.3 | 39.3 | 34 | 40.6 | — | — | 100 | 65.1 |
| Qwen3 Coder 30B A3B Instruct | 160K | — | — | — | — | — | 34.1 | 42.1 | — | 59.5 | — | — | — | |
| Qwen3 235B A22B Instruct 2507 | 262.1K | — | — | — | — | — | — | 60.6 | 59.4 | — | 96.2 | — | — | — |
| Gemini 2.5 Pro Preview 06-05 | 1.0M | — | — | — | — | — | — | 59.3 | 60.4 | — | 91.5 | — | — | — |
| Qwen3 30B A3B | 131.1K | — | — | — | — | — | — | 39.6 | 48.6 | — | 85.6 | — | — | — |
| Qwen3 8B | 131.1K | — | — | — | — | — | — | 31.6 | 45.8 | — | 82.5 | — | — | — |
| Qwen3 14B | 131.7K | — | — | — | — | — | — | 42 | 47.4 | — | 86.2 | — | — | — |
| Qwen3 32B | 131.1K | — | — | — | — | — | — | 45 | 51.6 | — | 88.4 | — | — | — |
| Qwen3 235B A22B | 131.1K | — | — | — | — | — | — | 51.1 | 54.8 | — | 88.5 | — | — | — |
| o4 Mini High | 200K | — | — | — | — | — | — | — | — | — | — | — | — | — |
| o3 Mini High | 200K | — | — | — | — | — | — | 56.6 | 56.6 | — | 92.2 | — | — | — |
| R1 Distill Llama 70B | 128K | — | — | — | — | — | 29 | 41.9 | — | 80.2 | — | — | — | |
| Mistral Large 2407 | 131.1K | — | — | — | — | — | — | 26.9 | 39.6 | — | 40.4 | — | — | — |
| Hermes 3 70B Instruct | 131.1K | — | — | — | — | — | — | 21 | 33.8 | — | 28.1 | — | — | — |
| GPT-4o (2024-08-06) | 128K | — | — | — | — | — | — | 32.4 | 27.5 | — | 45.6 | — | — | — |
| GPT-4o (2024-05-13) | 128K | — | — | — | — | — | — | 32.2 | 43.1 | — | 45.1 | — | — | — |
| GPT-3.5 Turbo (older v0613) | 4.1K | — | — | — | — | — | — | — | — | — | — | — | — | — |
