Pricing for light-speed inference.

Production speech inference with the lowest-cost, usage-based pricing.
No platform subscription.

Private

Dedicated Inference

Run your model on dedicated capacity, sized for your workload.

Model APIs Pricing

Instant access to Nari’s fastest, most affordable speech models through managed APIs.

ModelClassPriceAction
Nari Qwen3-TTSStandard
$5.00per 1M input characters
Try Model API
Nari Qwen3-TTS
$10.00per 1M input characters
Try Model API
Nari Qwen3-ASRStandard
$0.06per input audio hour
Try Model API
Nari Qwen3-ASR
$0.12per input audio hour
Try Model API

Top up freely.
Unlock more with $100+.

Top up as you go

$10~

  • 3 member seats
  • 2 concurrency per model

Unlock more forever

$100~in one payment

  • 10 member seats
  • 20 concurrency for each TTS and STT

Need more seats, concurrency, or volume discounts? Contact us.

FAQ

A single $100+ credit purchase in your organization unlocks 10 member seats and higher concurrency. Purchases under $100 don’t combine.