Skip to content
Sign in
TextNew

DeepSeek V4 Pro

Большая MoE-модель DeepSeek: рассуждения и код по низкой цене.

Try DeepSeek V4 ProPricing and top-ups

Price

Charged in Mesh tokens from your shared balance or subscription. The price is shown before you run.

Input text per 1,000 tokens
0.14 MT≈ ₽0.04
Model response per 1,000 tokens
2.66 MT≈ ₽0.80
Repeated input (cache) per 1,000 tokens
0.08 MT≈ ₽0.03
  • The plan discount (Pro, Ultra) applies to this price.

Capabilities

  • Function calling
  • JSON output
  • Reasoning
  • Streaming responses
  • Context 1M tokens

Via API

The model is available in the OpenAI-compatible API under the identifier: deepseek-v4-pro

API docs

Use cases

  • Answers to questions, explanations and translations
  • Copy for websites, emails and social media
  • Code: write it, explain it, find the bug
  • Document analysis and summaries of long texts
  • Agents and automation: call your own functions and APIs
  • Complex tasks: math, logic, planning

Similar models