Pareto Inference

Pareto Inference defaults to zero data retention for its standard inference API: prompts and outputs are not kept in persistent storage after a request finishes, and model training requires separate opt-in. Temporary in-memory prompt caching may be used; operational, security, account, and billing metadata may be retained. Upstream fallback is restricted to providers OpenRouter designates as zero data retention. See https://paretoinference.com/zdr for scope and exceptions.

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 Pareto Inference

Model name

Model	Context	Input /M	Output /M	Cache read /M	TTFT	Throughput
DeepSeek V4.1 Flash deepseek/deepseek-v4.1-flash:pareto	1.0M	$0.25/M	$1.01/M	$0.0050/M	N/A	N/A
DeepSeek V4.1 Flash (Thinking) deepseek/deepseek-v4.1-flash:thinking:pareto	1.0M	$0.25/M	$1.01/M	$0.0050/M	N/A	N/A
GLM 5.3 Flash z-ai/glm-5.3-flash:pareto	1.0M	$0.06/M	$0.21/M	$0.01/M	N/A	N/A

DeepSeek V4.1 Flash

deepseek/deepseek-v4.1-flash:pareto

Context

1.0M

Input /M

$0.25/M

Output /M

$1.01/M

Cache read /M

$0.0050/M

TTFT

N/A

Throughput

N/A

DeepSeek V4.1 Flash (Thinking)

deepseek/deepseek-v4.1-flash:thinking:pareto

Context

1.0M

Input /M

$0.25/M

Output /M

$1.01/M

Cache read /M

$0.0050/M

TTFT

N/A

Throughput

N/A

GLM 5.3 Flash

z-ai/glm-5.3-flash:pareto

Context

1.0M

Input /M

$0.06/M

Output /M

$0.21/M

Cache read /M

$0.01/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 Pareto Inference via the API

Call a model served by Pareto Inference 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": "deepseek/deepseek-v4.1-flash:pareto",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

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