Astorias

Astorias publishes zero retention and no training for inference prompts and outputs, with legal, abuse, and security exceptions. Account, billing, and operational metadata may be retained. Automatic prompt caching is supported; cross-request cache hits have been observed, but the published request-only KV lifetime wording does not establish a cache retention period. Inference location has not been verified.

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.

Models served by Astorias

Model name

Model	Context	Input /M	Output /M	Cache read /M	Cache hit	TTFT	Throughput
GLM 5.3 Flash NVFP4 z-ai/glm-5.3-flash:astorias	1.0M	$0.08/M	$0.26/M	$0.02/M	91.4%	1.3s	97.7 tps
Qwen3.8 27B FP8 qwen3.8-27b:astorias	262k	$0.11/M	$1.68/M	$0.05/M	N/A	N/A	N/A
Qwen3.8 27B Thinking FP8 qwen3.8-27b:thinking:astorias	262k	$0.11/M	$1.68/M	$0.05/M	N/A	N/A	N/A

GLM 5.3 Flash

NVFP4

z-ai/glm-5.3-flash:astorias

Context

1.0M

Input /M

$0.08/M

Output /M

$0.26/M

Cache read /M

$0.02/M

Cache hit

91.4%

TTFT

1.3s

Throughput

97.7 tps

Qwen3.8 27B

FP8

qwen3.8-27b:astorias

Context

262k

Input /M

$0.11/M

Output /M

$1.68/M

Cache read /M

$0.05/M

Cache hit

N/A

TTFT

N/A

Throughput

N/A

Qwen3.8 27B Thinking

FP8

qwen3.8-27b:thinking:astorias

Context

262k

Input /M

$0.11/M

Output /M

$1.68/M

Cache read /M

$0.05/M

Cache hit

N/A

TTFT

N/A

Throughput

N/A

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 Astorias via the API

Call a model served by Astorias using its displayed model ID, including the provider suffix when selection is supported.

curl https://nano-gpt.com/api/v1/chat/completions \
  -H "Authorization: Bearer $NANOGPT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "z-ai/glm-5.3-flash:astorias",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

For models with provider selection, append :astorias to the model ID, set "provider": "astorias" in the request body, or send an X-Provider: astorias 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.