deepseek-v4-flash

Overview

A flagship MoE large model with 1.6 trillion parameters and 49 billion activated parameters, natively supporting context lengths of up to one million tokens. Trained on a vast corpus of high-quality data, it excels in advanced mathematical reasoning, complex logical inference, specialized coding, and deep analysis of long-form text, making it well-suited for demanding applications such as cutting-edge research, sophisticated office workflows, and advanced AI agents.

• Model type: Text generation
• Input: Text
• Output: Text
• Endpoints: chat_completion / chat

Capabilities

• Near-flagship inference level: On simple Agent tasks, it performs on par with V4-Pro, covering the vast majority of daily intelligent agent and productivity scenarios.
• Extremely fast response: The activation parameters are only 13B, which is about 1/4 of that of V4 Pro. Its response speed is much faster than that of V4 Pro, and it supports millions of long contexts by default, maintaining stable low latency in long text scenarios.
• Ultimate cost-effectiveness: Approaching the upper limit of V4 Pro's capabilities with approximately 1/6 of the total parameter count, priced at less than one-tenth of the flagship model.

Code example

Call with deepseek-v4-flash :

curl https://www.starunion.net/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4-flash",
    "messages": [{ "role": "user", "content": "Hello" }]
  }'