DeepSeek: deepseek-v4-pro
DeepSeek V4 Pro is a large-scale Mixture-of-Experts model from DeepSeek with 1.6T total parameters and 49B activated parameters, supporting a 1M-token context window. It is designed for advanced reasoning, coding, and long-horizon agent workflows, with strong performance across knowledge, math, and software engineering benchmarks. Built on the same architecture as DeepSeek V4 Flash, it introduces a hybrid attention system for efficient long-context processing. Reasoning efforts high and xhigh are supported; xhigh maps to max reasoning. It is well suited for complex workloads such as full-codebase analysis, multi-step automation, and large-scale information synthesis, where both capability and efficiency are critical
Precios
Tiempo de actividad
Rendimiento
Uso y clasificación
Parámetros compatibles
Todos los proveedores = compatible en todos los upstreams que sirven este modelo. Algunos proveedores = depende del upstream que atienda la solicitud. Predeterminado = el valor enviado cuando no lo configuras.
| Parámetro | Proveedores | Predeterminado |
|---|---|---|
| frequency_penalty | Todos los proveedores | - |
| include_reasoning | Todos los proveedores | - |
| logit_bias | Algunos proveedores | - |
| logprobs | Algunos proveedores | - |
| max_completion_tokens | Algunos proveedores | - |
| max_tokens | Todos los proveedores | - |
| min_p | Algunos proveedores | - |
| presence_penalty | Todos los proveedores | - |
| reasoning | Todos los proveedores | - |
| reasoning_effort | Todos los proveedores | - |
| repetition_penalty | Algunos proveedores | - |
| response_format | Todos los proveedores | - |
| seed | Algunos proveedores | - |
| stop | Todos los proveedores | - |
| structured_outputs | Todos los proveedores | - |
| temperature | Todos los proveedores | 1 |
| tool_choice | Todos los proveedores | - |
| tools | Todos los proveedores | - |
| top_k | Todos los proveedores | - |
| top_logprobs | Algunos proveedores | - |
| top_p | Todos los proveedores | 1 |