Context windows and maximum output tokens across leading models
Machine-assisted translation of the archived article. Read the Chinese original.

| The author argues that increasing the output limit from 4K to 8K tokens can increase compute costs sharply. | |||||
| Context windows and maximum output tokens for major language models | |||||
| No. | Model | Provider | Context window | max output token | Notes |
| 1 | o1 and variants | openAI | 200k | 100k | The author’s top pick at publication |
| 2 | o1-mini and variants | OpenAi | 128k | 64k | |
| 3 | o1-preview and variants | openai | 128k | 32k | |
| 4 | GPT-4o and variants | openai | 128k | 4k | |
| 5 | GPT-4 and variants | openai | 8-32k | 4k | |
| 6 | GPT-4o mini and variants | openai | 128k | 4k | |
| 7 | Claude 3.5 and variants | openai | 200k | 4k | |
| 8 | Moonshot-v1-32K | Moonshot AI | 32k | 4k | |
| 9 | Moonshot-v1-128K | Moonshot AI | 128k | 4k | |
| 10 | deepseek v3 | DeepSeek | 64k | 8k | |
| 11 | Doubao-pro | ByteDance | 256k | 4k | |
| 12 | Doubao-lite-32k | ByteDance | 32k | 4k | |
| 13 | Doubao-lite-128k | ByteDance | 128k | 4k | |
| 14 | GLM-4 and variants | Zhipu AI | 8k | 4k | |
| 15 | GLM-Zero-Preview | Zhipu AI | 16k | 12k | Of the 16K total tokens, this model allows up to 12K tokens of output. |
| 16 | GLM-4-Plus | Zhipu AI | 128K | 4k | |
| 17 | hunyuan-large-longcontext | Tencent Hunyuan | 128k | 6k | |
| 18 | hunyuan-large | Tencent Hunyuan | 28k | 4k | |
| 19 | hunyuan-turbo and variants | Tencent Hunyuan | 28k | 4k | |
| 20 | Qwen-Turbo | Alibaba | 1000k | 4k | |
| 21 | Qwen-Long | Alibaba | 1000k | 4k | |
| 22 | Qwen-Plus | Alibaba | 128k | 4k | |
| 23 | Qwen-Max | Alibaba | 32k | 2k | |
