Streaming models

Requirements and limitations

Handle partial chunks, interrupted connections, and final usage separately. Streaming does not imply lower total latency.

Only active chat models with explicit support are included. Missing capability values remain unknown; top-level values take precedence over nested values. Image input also establishes vision support.

Example request

POST https://api.llm7.io/v1/chat/completions
Authorization: Bearer YOUR_LLM7_API_KEY
Content-Type: application/json

{
  "model": "DeepSeek-V4-Flash-0731",
  "messages": [
    {
      "role": "user",
      "content": "Hello!"
    }
  ],
  "stream": true
}

Explore related pages