DiscreteStack

DiscreteStack AD (Bulgaria) states that prompts and outputs are processed in volatile memory only, never persisted or used for training. Prompt caches are partitioned per customer and deleted after 15 minutes of inactivity or two hours. Content-free request metadata is kept up to 90 days, and all processing, including subprocessors, stays in the EU.

Models served by DiscreteStack

Model name

Model	Context	Input /M	Output /M	Cache read /M	TTFT	Throughput
GLM 5.3 Flash NVFP4 Unverified z-ai/glm-5.3-flash:discretestack	1.0M	$0.0525/M	$0.21/M	$0.0158/M	N/A	N/A

GLM 5.3 Flash

NVFP4

Unverified

z-ai/glm-5.3-flash:discretestack

Context

1.0M

Input /M

$0.0525/M

Output /M

$0.21/M

Cache read /M

$0.0158/M

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 DiscreteStack via the API

Call a model served by DiscreteStack 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:discretestack",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

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