mixgateway.io

Model Naming & Routing

1. Convention for the model field

The gateway uses a single model string to identify both the provider and the upstream model, separated by a colon:

  • Format: provider:modelId

Examples:

  • DeepSeek chat: deepseek:deepseek-chat
  • Calling a provider’s native model id directly (when supported): deepseek-chat
  • Another host naming the same logical model: AliyunBailian:deepseek-chat (example)
  • DeepSeek reasoning: deepseek:deepseek-reasoner
  • OpenAI GPT‑4.1 mini: openai:gpt-4.1-mini

Where:

  • provider — Identifier inside the gateway, e.g. deepseek, openai, siliconflow, AliyunBailian, etc.
  • modelId — The concrete model name as in that provider’s docs, e.g. deepseek-chat, deepseek-reasoner, gpt-4.1-mini.

You can sometimes omit the provider: prefix. For example, deepseek-chat alone may resolve via the gateway’s default mapping for that model id.

2. Routing behavior (transparent to callers)

When you call POST https://api.mixgateway.io/v1/chat/completions:

  1. The gateway parses provider and modelId from model (or applies defaults when you send only modelId).
  2. It selects the correct upstream endpoint and credentials.
  3. Before forwarding, it may rewrite model to the provider’s real id (e.g. deepseek-chat) and attach the right upstream authorization.

You do not need to maintain provider mappings in your app. If model is valid for your account, the gateway routes to the intended provider and model.

3. Example: migrating from DeepSeek to the gateway

Original (directly to DeepSeek):

curl https://api.deepseek.com/v1/chat/completions \
  -H "Authorization: Bearer {deepseek_api_key}" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-chat",
    "messages": [{"role":"user","content":"Hi"}]
  }'

Via mixgateway.io (same body, different base URL and key):

curl https://api.mixgateway.io/v1/chat/completions \
  -H "Authorization: Bearer {gateway_api_key}" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-chat",
    "messages": [{"role":"user","content":"Hi"}]
  }'

Typical changes:

  • Base URL — from the vendor host to https://api.mixgateway.io/v1 (no path segment beyond /v1 if that is how your deployment exposes the API).
  • Authorization — from the vendor API key to your gateway API key from the [console](/dashboard).

Fields like model, messages, temperature, max_tokens, etc. usually stay the same.

You may also use an explicit provider:

"model": "deepseek:deepseek-chat"

4. Supported providers & models

  • Use the public Pricing page on the site for published rates and high-level model listings.
  • In the [console](/dashboard), confirm which providers and modelId values are enabled for your account and review billing rules before production traffic.

See also