Pick one of five music takes: the audition costs $0.625 on Sume
Lyria 3.5 results vary per call, so auditioning is normal. On Sume's fixed $0.125 Music Router price, five takes cost $0.625. How to run and keep them.

Five takes from the Sume Music Router cost $0.625, because every accepted generation is a fixed $0.125 whichever engine runs. The Music Router docs say every model charges the fixed Music price per audio generation, and the pricing page lists Music at $0.125 per audio. Neither mentions seeds or iterative edits, so an audition is simply more calls with the same brief. Both pages were read on 2026-10-03.
What does an audition cost?
Price does not vary with prompt length or the optional image_url, per the Music 1.0 docs, so the table is just multiplication.
| Takes | List-rate cost |
|---|---|
| 1 | $0.125 |
| 3 | $0.375 |
| 5 | $0.625 |
| 10 | $1.25 |
How do I run five takes safely?
Send five requests to POST /v1/music-router/generate, each with its own Idempotency-Key such as bed-take-1 to bed-take-5. Reusing one key would return the first job, and reusing it with a changed payload returns 409 idempotency_conflict, so one key per take. Submit with mode: "async" and poll GET /v1/jobs/:id/status; the jobs docs say not to resubmit a paid request just because a local timeout fired.
How do I judge them?
- Vary one thing between takes, such as tempo or the lead instrument, so you learn what the prompt controls.
- Read
result.artifacts[]wheretypeisaudio, and keep themedia.sume.comURL; raw provider URLs are not public outputs. - Check
job.request.routed_modelto see which engine ran.sume/music-autopicks Lyria 3.5 today.
What is not guaranteed?
Nothing in the docs promises that takes differ much, or that any one fits your picture. The brief is a creative direction, not an output setting, so verify the audio by ear before it goes under a video.
Sources
Related posts
More in Pricing
- Pictory AI credits in dollars: Starter $0.29, Professional $0.118
Pictory's AI credits cost $0.29 each on Starter monthly, $0.118 on Professional and $0.083 on Team. Annual billing brings them to $0.25, $0.035 and $0.05.
- Q4 2026 AI video budget template: lean, standard, heavy by month
A Q4 2026 template with three scenarios by month, priced from the Sume rate card: screens, Seedance 2.5 heroes, Wan 3.0 finals and captions.
- Qwen-Image at $0.02 per megapixel vs a flat price per image
fal bills Qwen-Image per megapixel, rounded up, so a 1024x768 render counts as one. What that means when Sume prices the same model per image.
- Real-time AI avatar pricing: live minutes vs one render
Live avatars bill by conversation minute and concurrent stream; a rendered clip is paid once and watched by anyone. Tavus plan numbers and the break-even.
Written by Sume