ModelsReference
Model catalog
Every model, its provider, category, and availability.
By the end: You can pick a model by capability, price, and context size.
The catalog lists every model this deployment can serve, with provider, category, context window, and availability. The live list is authoritative.
List models
curl http://localhost:3000/v1/models \
-H "Authorization: Bearer $ROUTEFORGE_API_KEY"Categories
| Field | Value | Notes |
|---|---|---|
| Cloud | OpenAI, Anthropic, Google, xAI | General-purpose text and reasoning models. |
| Chinese | DeepSeek, Zhipu, Moonshot, Tencent, MiniMax, Alibaba | Strong cost-to-quality options, often with large context. |
| Image | Multiple | Text-to-image. Priced per image. |
| Video | Multiple | Text-to-video. Priced per clip. |
One key, every family
You do not need separate keys for different providers. A single RouteForge key reaches every model in the catalog; the gateway selects the correct upstream service per request based on the model you name.
Availability
Available means the upstream service currently lists the model. Disabled means an administrator turned it off. Unavailable means it is in the catalog but the upstream no longer lists it — a request will be refused with model_not_found.