Lyria 3.5 vs Suno v6 for ad background music
Lyria 3.5 is an API at $0.08 a song with SynthID. Suno v6 is a subscription with commercial rights only on paid plans. Pick by workflow, then check terms.

For ad background music produced by a pipeline, Lyria 3.5 is the better fit: it is a per-song API at $0.08 with documented output specs. Suno v6 fits a human producer working in an app, and its commercial-use rights depend on your paid plan. Neither page I read settles brand-licensing questions for you, so treat the table below as a workflow comparison, not legal advice.
Side by side
Everything in the table is from the vendor's own page, except the Suno v6 training-data row, which comes from TechCrunch.
| Lyria 3.5 (Google) | Suno | |
|---|---|---|
| How you buy | Gemini API, $0.08 per song | Subscription: Free, Pro $8/mo, Premier $24/mo |
| Output | 44.1 kHz stereo MP3 or WAV | Credits and monthly downloads per plan |
| Length | Full songs of a couple of minutes, set by prompt | Not covered in the pages I read |
| Inputs | Text plus up to 10 images; custom lyrics; instrumental-only | Not covered in the pages I read |
| Provenance | SynthID audio watermark on all output | TechCrunch: v6 models trained on licensed music; watermarking added |
| Blocked | Prompts for specific artist voices or copyrighted lyrics | Not covered in the pages I read |
| Commercial use | Not addressed on the pages I read | Free: no commercial rights. Pro and Premier: commercial use |
| Editing | Single-turn only, outputs vary between identical calls | Edit-by-text and mashups reported by TechCrunch |
Where Lyria 3.5 wins for ads
Ad work is repetitive: one brief, many cut lengths and moods. A per-song API charges by the song, so you can generate twenty candidates for $1.60 at Google's list ($0.08 x 20) and pick one. Because Lyria takes timestamps like [0:00-0:10] and section tags in the prompt, you can ask for a sting at the start and a resolve at the end of a 15-second spot.
The block on artist voices and copyrighted lyrics is a feature here. A brand cannot ship a track that sounds like a named singer, and the model refuses to try.
Where Suno wins
Suno is built for iteration by a person. Pro lists 2,500 credits and 20 downloads a month for $8; Premier lists 10,000 credits and 60 downloads for $24, plus Suno Studio and 30-minute uploads. If a creative director wants to audition many options and download a few, the app suits that better than an API.
- Free tier: 50 credits a day on free models, no commercial rights. Do not use free-tier tracks in ads.
- Download caps matter: 20 downloads a month on Pro means 20 deliverable tracks, not 20 attempts.
- TechCrunch reports the v6 models were not trained on the data used for earlier versions, older models are to be retired, and Sony and UMG suits are still active.
What Sume offers
Sume ships Lyria 3.5, not Suno. Music 1.0 runs on Lyria 3.5 through POST /v1/music-router/generate, at a fixed $0.125 per accepted generation. It rejects duration and non-empty negative_prompt, so you steer length and exclusions in the prompt. If you need Suno's app workflow, you use Suno directly; Sume does not resell it.
Before an ad runs, have your legal team read the current terms of whichever service you use. The pages above tell you what the tools do, not what you may do with the output.
Sources
Related posts
More in Comparisons
- MAI-Transcribe-2 vs Sume STT 1.0: keywords, diarization, price
MAI-Transcribe-2: keyword biasing, diarization, $0.10/hr limited-time. Sume STT 1.0: $0.01 per minute ($0.60/hr), word timings always on, no keyword field.
- MAI streaming regions vs Sume: no region field in the STT request
MAI-Transcribe-2-Streaming lists swedencentral, centralus, southindia (eastus2 soon). Sume's STT request has no region field, so ask support about residency.
- MAI-Voice-2.1 languages vs Sume voice tags: 14 shared, 9 and 2 apart
Microsoft lists 23 TTS languages; Sume's voice library tags 16. Fourteen overlap, nine are MAI-only, two (Japanese, Tagalog) are Sume-only. The full split.
- Flash TTS '55% faster inference' vs the end-to-end time of a TTS job
Microsoft quotes 55% faster inference and 150 ms end to end for MAI-Voice-2.1-Flash. Sume TTS is an async job; here is what you can actually time.
Written by Sume