NVIDIA: Nemotron 3.5 Transcribe
Speech to textNemotron 3.5 Transcribe is an open-weight speech recognition model from NVIDIA, a 30B mixture-of-experts with about 4B active parameters.
- Price
- 361 Toman a minute of audio
- Endpoint
- POST /v1/audio/transcriptions
- Published
- Oct 29, 2026
- Offers
- With segment and word times
- Maker
- NVIDIA
Free plan prices, in Toman at today's rate. Plans and prices
Price on each plan
Per millisecond of audio heard, shown a minute at a time, in Toman at today's rate; never below what the provider charges us plus your plan's margin.
| Plan | Price |
|---|---|
| Free | 361 Toman a minute of audio |
| Pro | 346 Toman a minute of audio |
| Max | 331 Toman a minute of audio |
Call it
curl https://ai.fluxcdn.cloud/api/v1/audio/transcriptions -H "Authorization: Bearer $FLUXCDN_API_KEY" \
-F file=@voice.mp3 -F model=nvidia/nemotron-3.5-transcribe -F language=fa \
-F response_format=srt
