A 40-second AI music sting: ElevenLabs $0.10 beats Sume's $0.125 flat
ElevenLabs lists music at $0.15 per minute, so 40 seconds is $0.10. Sume Music charges a flat $0.125 per generation, which wins only above 50 seconds of audio.

For a 40-second sting, ElevenLabs' listed music price of $0.15 per minute works out to $0.10, which is less than Sume's flat $0.125 per generation. Sume becomes cheaper only when a track runs longer than 50 seconds, because $0.125 / $0.15 per minute = 0.833 minutes. At two minutes, ElevenLabs is $0.30 against Sume's $0.125.
Prices were read on ElevenLabs' API pricing page and Sume's catalog on 2026-10-09.
Break-even by length
Sume Music Router charges the same fixed price on every catalog model, and Music 1.0 requests resolve through the router at the same fee. The price does not vary by prompt or image input. ElevenLabs shows a per-minute rate, so its price rises with length.
| Track length | ElevenLabs at $0.15/min | Sume at $0.125 per generation |
|---|---|---|
| 15 seconds | $0.0375 | $0.125 |
| 40 seconds | $0.10 | $0.125 |
| 50 seconds | $0.125 | $0.125 |
| 90 seconds | $0.225 | $0.125 |
| 180 seconds | $0.45 | $0.125 |
The catch with a flat fee
Sume's music call has no duration field. The docs say Music 1.0 and the router reject duration and duration_seconds, and that you control length in the prompt, for example by asking for a track of 30 seconds or using section markers like [0:00-0:30]. The prompt guides length but is not a guarantee, so measure what comes back. A sting that arrives at 55 seconds instead of 40 would still need a trim before you use it.
If you generate several 40-second options and keep one, the flat fee also multiplies: three attempts at $0.125 is $0.375. Under ElevenLabs' per-minute price, three 40-second attempts are $0.30 on the same arithmetic.
Choosing for a project
Short, repeated stings favor a per-minute price. Full-length beds and songs favor a flat one. If your video render already runs through Sume, the Music output arrives as a Sume-hosted audio file that the timeline can use as a soundtrack, which saves an upload step. That convenience is worth a few cents only if the numbers above fall your way.
- Under 50 seconds: ElevenLabs is cheaper per attempt on these prices.
- Over 50 seconds: Sume's flat fee is cheaper.
- Always check the delivered length, since Sume cannot set it by parameter.
Why the difference is small either way
At these lengths the gap is a couple of cents per track. A 15-second logo sting is $0.0375 on a per-minute price and $0.125 flat, so the flat fee is more than three times higher for that one file, but it is still twelve and a half cents. Time spent choosing the cheapest option for a single sting is worth more than the saving.
The gap matters for volume. A game studio that wants 200 short effects pays $25.00 at Sume's flat fee ($0.125 x 200) against $7.50 at a per-minute rate for 15-second files ($0.0375 x 200). For beds of two to three minutes the order reverses, and the flat fee is the cheaper one.
Sources
Related posts
More in Pricing
- 402 insufficient_credits on a 30-second 1080p video: balance needed
Sume reserves the full estimate at submit. A 30-second Seedance 2.5 1080p job needs $42.65 spendable before it is accepted; four other jobs compared.
- 42 museum audio-guide stops of 380 characters: MAI vs Sume TTS bill
A 42-stop audio guide at 380 characters a stop is 15,960 characters. MAI-Voice-2.1 costs $0.35, Flash $0.24, Sume TTS $0.76. The arithmetic and the catch.
- 45-minute video render on Sume: the 30-minute cap and a $4.50 split
Timeline 1.0 renders cost $0.10 per output minute, rounded up, with a 30-minute maximum. A 45-minute video renders as 30 + 15 minutes for $4.50.
- 45-second Short from three MiniMax H3 Max clips at 1080p: $9.10
Three 15-second 9:16 MiniMax H3 Max jobs at 1080p cost $9.00, plus $0.10 to join them in Timeline: $9.10 for a 45-second YouTube Short.
Written by Sume