Token counts are estimates based on each model's tokenizer characteristics; actual values may differ by 5-15%
模型
Input Cost
Output Cost
单价 ($/MTok)
Costs are estimated based on official platform pricing, for reference only
Token Optimization Tips
1. 削除多余スペース和空行，减少无意义 Token 2. Use concise expressions instead of verbose descriptions to reduce prompt length 3. Code text typically has higher token density; simplify comments and variable names 4. 中文テキスト在 Gemini / Qwen 中 Token 效率更高 5. English text varies less across models; Chinese varies more significantly