CanalAPI
Developer Guide

Models

How CanalAPI exposes hundreds of models behind a single endpoint.

CanalAPI aggregates models from multiple providers behind a single OpenAI-compatible endpoint. You select the model in each request via the model field.

Listing models

The console Models page is the authoritative source for the current model list. It shows for each model:

  • The canonical model name to use in the model field.
  • Provider and family.
  • Context window (max tokens).
  • Capabilities (text, vision, tools, reasoning, etc.).
  • Pricing.

Always refer to the console for the latest names and prices. The list evolves as providers ship new models.

Selecting a model

Set the model field in the request body:

{
  "model": "gpt-4o-mini",
  "messages": [{"role": "user", "content": "Hello"}]
}

Common families include:

  • OpenAI — GPT-4o and the o-series reasoning models.
  • Anthropic — Claude.
  • Google — Gemini.
  • DeepSeek, Moonshot/Kimi, Zhipu, Alibaba/Qwen — leading Chinese open and proprietary models.

The exact model names available to your account depend on regional availability and your plan.

Choosing the right model

A practical rule of thumb:

ScenarioSuggested
High-volume classification, summarization, routingSmaller, faster, cheaper model
Conversational assistant, general-purpose useMid-tier flagship model
Complex reasoning, code synthesisTop-tier or reasoning model
Vision inputA model with documented vision capability
Long context (100k+ tokens)A model whose context window covers your need

Compatibility notes

  • All listed models accept the OpenAI Chat Completions format.
  • Provider-specific features that are not represented in the OpenAI format may be exposed as additional request fields. These are documented in the console for each model where applicable.
  • Behavior, output quality, and pricing differ across models even for nominally similar tasks. Benchmark with your own evaluation set before committing to a specific model in production.

On this page