Fact · Transcription
$0.12US dollars per hour of audio
Nari Labs' pricing page lists its Nari Qwen3-ASR Fast speech recognition model at $0.12 per input audio hour.
Quote found word for word on the live source, last checked October 21, 2026
In the source's own words
“Nari Qwen3-ASR Fast #1 in TTFS on Coval $0.12 per input audio hour”
Model APIs Pricing | Nari Labs, Nari Labs, 2026, Model APIs Pricing table.
Open the original- What it measures
- The Model APIs Pricing table lists the Fast class at twice the Standard price for the same unit, per input audio hour. Models are paid with prepaid credits. The page's Fast class is the low-latency option, and the home page lists it with a sub 40 millisecond server-side time to first segment. The page does not say whether the price is for batch or streaming use. This row is recorded as streaming, on the same basis as the Standard row.
- What it does not show
- The mode is inferred, not stated: the price page names no batch or streaming split, and the documentation's only Speech to Text guide is realtime transcription. The '#1 in TTFS on Coval' label is the vendor's own ranking claim and is not part of the price.
In the data
A row of Speech-to-Text API Pricing: Cost per Hour of Audio, as Nari Qwen3-ASR Fast: 0.12 US dollars per hour of audio.
- Provider
- Nari Labs
- Product
- Nari Qwen3-ASR Fast
- Mode
- Streaming (real time)
- Tier
- Top up as you go
- As of
- 2026-10
- Note
- The page states $0.12 per input audio hour. No conversion.
Source
Used to calculate
- In the speech-to-text and transcription prices dataset, the 36 published figures for “Streaming (real time)” have a median of 0.39 US dollars per hour of audio, ranging from 0.01 to 2.
- The speech-to-text and transcription prices dataset holds 117 data points, each a published figure traced to its source.
- The speech-to-text and transcription prices dataset draws on 43 distinct sources.