Luma Whisper v3 at 1 credit a minute: transcription cost per hour
An hour of Whisper v3 on Luma is 60 credits, $0.10 to $0.18 by plan. ElevenLabs Scribe v2 lists $0.22 an hour; Sume's speech-to-text is $0.60 an hour.

An hour of Whisper v3 on Luma costs 60 credits, which is $0.18 on Plus, $0.135 on Pro and $0.12 on Ultra at monthly prices, or $0.15, $0.1125 and $0.10 billed annually. ElevenLabs Scribe v2 is listed at $0.22 an hour. Sume's speech-to-text is $0.01 per audio minute, or $0.60 an hour, so for plain transcription Luma and ElevenLabs are cheaper per hour than Sume.
That is the honest ordering on price alone. What Sume's rate buys is a per-job charge with no plan, which matters mostly when transcription is a small side task.
What does each option charge per hour of audio?
Luma's pricing page lists Plus at $30 a month ($25 annual) for 10,000 credits, Pro at $90 ($75) for 40,000, and Ultra at $300 ($250) for 150,000, with Whisper v3 at 1 credit per minute. ElevenLabs lists Scribe v2 at $0.22 per hour on its API pricing page. Sume lists speech-to-text at $0.01 per audio minute.
| Option | Monthly price | Annual price | Notes |
|---|---|---|---|
| Luma Plus, Whisper v3 | $0.18 | $0.15 | 10,000 credits cover 166 hours |
| Luma Pro, Whisper v3 | $0.135 | $0.1125 | 40,000 credits cover 666 hours |
| Luma Ultra, Whisper v3 | $0.12 | $0.10 | 150,000 credits cover 2,500 hours |
| ElevenLabs Scribe v2 API | $0.22 | n/a | Per-hour API rate |
| Sume speech-to-text | $0.60 | n/a | $0.01 per minute, up to 10 minutes per job |
What is the 10-minute limit on Sume's side?
Sume's speech-to-text takes up to 10 minutes of audio per job. A one-hour recording therefore runs as six jobs of 10 minutes, still $0.60 in total, but you split and merge the transcript yourself. The Luma and ElevenLabs pages used here do not state a per-file cap, so I make no claim about theirs; check before you rely on a long recording.
When does the gap matter?
At one hour a day, the difference between Sume ($18 a month at 30 hours) and Luma Plus ($5.40) is under $13 a month. Over a library of thousands of hours it is a real line item, and Luma or ElevenLabs is the cheaper route on these pages. If transcription feeds a video job on the same wallet, the extra cents buy one balance and one API.
ElevenLabs charges add-ons for entity detection and key terms on Scribe; the details are in ElevenLabs Scribe v2 pricing per hour. To see how big a Luma credit pool is in other units, read Luma Plus, Pro and Ultra credit pools.
How do I compare my own volume?
Multiply your monthly audio hours by each hourly figure in the table. If one option wins by under the cost of a coffee, choose on workflow rather than price.
Sources
Related posts
More in Pricing
- Does Lyria 3.5 on the paid Gemini API train on my prompts?
Google's Gemini pricing page lists data use as No on the paid tier for Lyria models and Yes on the free tier. Lyria 3.5 has no free tier.
- MiniMax H3: 2K by default there, 768p by default on Sume
MiniMax's page says H3 defaults to 2K; Sume's default is 768p, and 2K bills $0.1625 per second versus $0.075. Per-second and 10 s costs by resolution.
- MiniMax H3 Max price per second on Sume: 480p, 768p, 1080p vs Wan
H3 Max lists $0.05, $0.08 and $0.16 a second at 480p, 768p and 1080p. On Sume that is $0.0625, $0.10 and $0.20. Compared with Wan 3.0 at each step.
- Four edit variants in one call: n limits and cost per Sume image model
Most Sume image models accept n up to 4, Grok Imagine only 1. You pay cost_usd times n, so four variants of an edit cost four edits.
Written by Sume