DeepSeek-V4-Pro hybrid model with both non-thinking and thinking (default) modes. · Succeeded by deepseek-v4-pro-2026-07-31
function_callinghttps://api.deepseek.com/v1/chat/completionshttps://api.deepseek.com/v1/completions| Date | Input | Output | Cache read | Change |
|---|---|---|---|---|
| Apr 25, 2026 | $0.14 | $0.28 | $0.028 | — |
| Apr 26, 2026da3af4a | $0.14 | $0.28 | $0.003 | ↑0.0% |
| May 11, 202608f0ffe | $0.435 | $0.87 | $0.004 | ↑211% |
Input is ↑211% since Apr 25, 2026, from $0.14 to $0.435 per 1M tokens.
| Model | Context | Max out | Pricing |
|---|---|---|---|
| DeepSeek V4-Flash | 1M | 384K | $0.14$0.28 |
| DeepSeek V3.2 (Chat) | 128K | 8K | $0.28$0.42 |
| DeepSeek V3.2 (Reasoner) | 128K | 64K | $0.28$0.42 |
| DeepSeek V4-Pro | 1M | 384K | $0.44$0.87 |
/v1/models/deepseek/deepseek-v4-pro-2026-04-24DeepSeek-V4-Pro hybrid model with both non-thinking and thinking (default) modes.