Overview
deepseek-v4-pro is DeepSeek’s model on the Reasoning shelf. Call it through Nebutra Router’s OpenAI-compatible API, with no separate contract for each upstream. It is on sale and routed now.
Capabilities
- One endpoint — One API key and base URL cover every provider; switch models by changing the model id.
- Pay per use — Input —/1M · output —/1M (list price; your invoice is the final figure).
- Context — A public context window of about 1M, suited to long conversations and documents (the model’s own limit applies).
- Reasoning — Suited to planning, code and multi-step tool use.