- Context
- 512k
- Input /M
- $0.173/M
- Output /M
- $0.693/M
- Cache read /M
- $0.0115/M
- Cache write /M
- $0.173/M
- Cache hit
- 87.2%
- TTFT
- 2.6s
- Throughput
- 115.4 tps
Sable
Sable’s privacy policy (updated 13 Nov 2026) states that prompts and outputs are never logged, stored or used for training; they exist on its servers only while a request is processed. NanoGPT classifies Sable as Zero Data Retention on these public terms. Sable does keep a short-lived numerical prompt cache, usually discarded within a couple of hours to a day, and its usage records keep keyed one-way hashes of prompt beginnings.
Privacy can differ by model. See each model’s label.
Models served by Sable
- Context
- 512k
- Input /M
- $0.173/M
- Output /M
- $0.693/M
- Cache read /M
- $0.0115/M
- Cache write /M
- $0.173/M
- Cache hit
- 97.7%
- TTFT
- 2.4s
- Throughput
- 129.9 tps
Prices are per million tokens. TTFT is the observed time to first token. 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 Sable via the API
Call a model served by Sable using its displayed model ID, including the provider suffix when selection is supported.
cURL
curl https://api.nano-gpt.com/api/v1/chat/completions \
-H "Authorization: Bearer $NANOGPT_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "pqhaz/DeepSeek-V4.1-Flash-1017-heretic:sable",
"messages": [{"role": "user", "content": "Hello"}]
}'For models with provider selection, append :sable to the model ID, set "provider": "sable" in the request body, or send an X-Provider: sable 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.
See the API documentation for provider preferences, price-aware routing, and error behavior.