One podcast episode's audio bill: MAI, Lyria and transcript vs Sume

For one episode with a 5,000-character narrated intro, one music track and a 60-minute transcript, the vendor list prices add up to $0.695: MAI-Voice-2.1-Flash $0.075, Lyria 3.5 $0.08 and MAI-Transcribe-2-Streaming at its $0.54 introductory hourly price. On Sume the same three jobs cost $0.9625.

About the transcribe price

Microsoft's $0.54 per hour is an introductory price through the end of the year, per its announcement. Sume's speech-to-text is $0.01 per audio minute, prorated per second, documented in the Video Inspect docs.

What the gap buys

The difference is under 27 cents an episode, so over a 52-episode year it is about $13.9. Choose on workflow: Sume keeps the voice, music and transcript behind one balance and one rate card, and the Music Router is covered in its docs.

Recompute it in Python

The three line items, summed. The numbers are the ones in the table above.

sume = 5000 * 0.0475 / 1000 + 0.125 + 60 * 0.01
vendor = 5000 * 15 / 1e6 + 0.08 + 0.54
print(round(sume, 4), round(vendor, 3))