DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes
Details
| Field | Value |
|---|---|
| id | accounts/fireworks/models/deepseek-v4-pro-0813 |
| family | deepseek-thinking |
| modality | text->text |
| context_length | 1000000 |
| max_output_tokens | 384000 |
| release_date | 2026-08-12 |
| is_open_weights | yes |
| supports_tool_call | yes |
| supports_reasoning | yes |
| supports_structured_output | yes |
| supports_attachment | no |
| input_modalities | text |
| output_modalities | text |