MiniMax API: endpoints, authentication, models and limits
Connection details
| Base URL | https://api.minimax.io/v1 |
| Auth | Bearer token via MINIMAX_API_KEY |
| OpenAI SDK compatible | Yes |
| SDKs | OpenAI SDK, raw HTTP |
Models and token pricing
| Model | Model ID | Context | In / 1M | Out / 1M |
|---|---|---|---|---|
| MiniMax M3 | minimax-m3 | 1000K | $0.600 | $2.40 |
What is distinctive about this API
- Text, speech and video generation sit behind one account, which is unusual — most providers cover a single modality well.
- Long-context text models plus a strong audio stack make it a fit for media-heavy pipelines.
Constraints worth knowing first
- Video and audio endpoints are asynchronous and job-based; they do not follow the synchronous chat-completions shape.
- International and mainland endpoints differ, as with other China-headquartered providers.
Frequently asked
- What is the MiniMax API base URL?
- https://api.minimax.io/v1. Authenticate with a bearer token, conventionally read from the MINIMAX_API_KEY environment variable.
- Is the MiniMax API compatible with the OpenAI SDK?
- Yes. The chat completions surface is wire-compatible, so the official OpenAI client libraries work by pointing base_url at https://api.minimax.io/v1 and supplying a MINIMAX_API_KEY. No client library change is needed.
- How do I get a MiniMax API key?
- Create an account with the provider and generate a key from the developer console, then expose it to your application as MINIMAX_API_KEY. Keys are secrets: keep them server-side, never in client bundles or committed config.
- Which models does the MiniMax API expose?
- MiniMax M3. Context windows run to 1000K tokens.
Other provider APIs
MiniMax publishes official documentation at https://www.minimax.io/platform/document. Pricing and limits change; verify against the provider before committing to a budget.