Appendix A: Model Catalog
Available models depend on your token's group and platform configuration. Use these methods to get the current list.
In the console
Open the model catalog to check model IDs, prices, capabilities, and supported protocols. {baseUrl} is your platform API root without a trailing / or /v1.
Through the API
curl "{baseUrl}/v1/models" \
-H "Authorization: Bearer {apiKey}"
Replace {apiKey} with an sk- model API key created in the console. Use the same key as your target tool, and copy the exact data[].id value from the response. Model IDs are case-sensitive.
Choose by task
| Task | Capabilities to check |
|---|---|
| Chat and translation | Chat completions, context length, and price |
| Claude Code | Anthropic Messages protocol |
| Codex CLI | Responses API and tool calling |
| Knowledge base | Embeddings and, if needed, reranking |
| Images and video | Generation endpoint and synchronous or asynchronous response format |
The model IDs in these tutorials are examples, not a guarantee of access. For model not found, check token groups and troubleshooting.
MiniMax H3 family
The MiniMax H3 integration guide covers H3/H3-Max video, 2K upgrades, Context-IR, and their token model permissions. Upgrade and IR requests still use MiniMax-H3 in the body; tokens need MiniMax-H3-Regeneration and MiniMax-H3-Context-IR permissions respectively. A listed model does not imply support for every API format; use the corresponding native /v2 endpoint.