Open models, without the refusals.

Sable is an inference API that works with the OpenAI and Anthropic formats your tools already use. Prompts and outputs are never logged, stored or used for training. The first model available is DeepSeek V4.1 Flash.

Create an account Documentation

Request

POST /v1/chat/completions
{
  "model": "deepseek-v4.1-flash-1017",
  "stream": true,
  "messages": [{
    "role": "user",
    "content": "This reads the whole log into memory. Make it stream line by line."
  }]
}

Response, streamed