List models
GET /v1/models uses the OpenAI list shape (object: "list", data array of { id, object, created, owned_by, tier }).
Only GET /v1/models (no extra path segments) is public in middleware. Without credentials the resolver uses plan free for filtering. With X-API-Key or Authorization: Bearer (rk_ or access JWT), the plan comes from the key or token so the list matches what you can call.
curl -sS https://api.routing.run/v1/models
curl -sS https://api.routing.run/v1/models \
-H "X-API-Key: ${ROUTING_RUN_API_KEY}"
Get model details
GET /v1/models/ requires middleware auth (same headers as chat). Unknown id → {"error":"Model not found"} (JSON).
curl -sS "https://api.routing.run/v1/models/route/deepseek-v3.2" \
-H "X-API-Key: ${ROUTING_RUN_API_KEY}"
id, object, created, owned_by, tier (heuristic lite vs premium from config buckets), allowed (bool for your resolved plan), and providers (the ordered upstream list from config).
Available models
This table lists the published chat model IDs. Authoritative access for your workspace is whatever GET /v1/models returns when authenticated with your rk_ key (your plan may expose only a subset).
Chat models
| Model | Source | Description |
|---|
route/deepseek-r1 | DeepSeek | DeepSeek R1 reasoning |
route/deepseek-v3.2 | DeepSeek | DeepSeek V3.2 chat |
route/deepseek-v3.2-speciale | DeepSeek | DeepSeek V3.2 Speciale |
route/glm-4.7 | Zhipu | GLM-4.7 |
route/glm-4.7-flash | Zhipu | GLM-4.7 Flash |
route/glm-5 | Zhipu | GLM-5 |
route/glm-5-highspeed | Zhipu | GLM-5 Highspeed |
route/glm-5.1 | Zhipu | GLM-5.1 |
route/glm-5.1-precision | Zhipu | GLM-5.1 Precision |
route/kimi-k2.5 | Moonshot | Kimi K2.5 |
route/kimi-k2.5-highspeed | Moonshot | Kimi K2.5 Highspeed |
route/minimax-m2.5 | MiniMax | MiniMax M2.5 |
route/minimax-m2.7 | MiniMax | MiniMax M2.7 |
route/minimax-m2.7-highspeed | MiniMax | MiniMax M2.7 Highspeed |
route/qwen3.5-9b | Qwen | Qwen3.5 9B |
route/qwen3.5-397b-a17b | Qwen | Qwen3.5 397B A17B |
route/qwen3.5-plus | Qwen | Qwen3.5 Plus |
route/qwen3.6-plus | Qwen | Qwen3.6 Plus |
route/qwen3.6-plus-preview | Qwen | Qwen3.6 Plus Preview |
Pricing
GET /v1/pricing is not on the anonymous path list — pass X-API-Key or Authorization: Bearer so middleware lets the request through. The handler returns static config (same table for any authenticated caller):
curl -sS https://api.routing.run/v1/pricing \
-H "X-API-Key: ${ROUTING_RUN_API_KEY}"
{
"models": [
{
"model": "route/deepseek-v3.2",
"display_name": "DeepSeek V3.2",
"tier": "free",
"input_per_million": 0.0,
"output_per_million": 0.0
}
]
}
model_pricing config; see dashboard for what you are actually charged after multipliers.
Inference usage is metered per request. Daily caps and credits depend on your plan tier — see the dashboard for current limits.
Image generation
| Model | Source | Description |
|---|
route/minimax-image-1 | MiniMax | MiniMax Image Generation |
