Important Update: DeepSeek API Upgraded to DeepSeek-V3.1 & Pricing Adjustment
 |
| Dear Developers, We are pleased to announce that DeepSeek API has been fully upgraded to our new premier model: DeepSeek-V3.1. Key enhancements include:Unified Dual-Mode Support DeepSeek-V3.1 combines reasoning (deepseek-reasoner) and non-reasoning (deepseek-chat) capabilities in a single model. Legacy models will be retired.Performance Upgrades Significant improvements in Agent tasks, reasoning efficiency, and 128K-token context handling.The Strict-mode for Function Calling Beta API now supports strict-mode for function calling, ensuring the output always complies with the defined JSON schema. Anthropic API Compatibility The API is now compatible with Anthropic API, allowing seamless integration with Claude Code. |
| Pricing Adjustment (Effective Sept. 4, 2025, 16:00 UTC): |
| Input tokens: $0.07/million (cache hit) / $0.56/million (cache miss) Output tokens: $1.12/million Off-peak discounts discontinued simultaneouslyCurrent pricing remains until the effective date. Plan usage accordingly. |
| For more details, please visit DeepSeek API Documentation. |
| Thank you for your continued trust in DeepSeek.Best regards,The DeepSeek TeamAugust 20, 2025 |
| © 2025 DeepSeek. All Rights Reserved. |