Kitani
Data handling can vary by model and deployment. The privacy badge above shows the more conservative of the provider-wide fallback and the policies for routes on this page; each model page shows the policy for that specific route.
NanoGPT requires zero data retention for Kitani routes serving Qwen 3.8 27B, GLM 5.3, Kimi K3, DeepSeek V4.1 Flash, and GLM 5.3 Flash Uncensored, including thinking variants. Prompts and outputs are not retained or used for training; usage and billing metadata are retained. Kitani documents a 503 response when no qualified ZDR deployment is available. Its catalog reports encrypted inference transport (TLS) for the reviewed DeepSeek V4.1 Flash and uncensored routes. This is not a provider-wide ZDR commitment.
Models served by Kitani
Prices are per million tokens. Selectable routes include the applicable provider-selection markup; fixed routes show the standard model price. Availability and pricing refresh continuously; each model page shows the live provider comparison.
Using Kitani via the API
For models with provider selection, append :kitani to the model ID, set "provider": "kitani" in the request body, or send an X-Provider: kitani header. Explicit provider selection adds a route-specific markup over that provider's base price; the prices in the table above already include it. Models without provider selection use the displayed model ID without a provider suffix or selection surcharge.
curl https://nano-gpt.com/api/v1/chat/completions \
-H "Authorization: Bearer $NANOGPT_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "lumen-stealth:kitani",
"messages": [{"role": "user", "content": "Hello"}]
}'See the API documentation for provider preferences, price-aware routing, and error behavior.