Google Lyria 3.5 vs Suno V6: Which Gives Creators More Control?

AI Music Tool Comparison · Updated September 12, 2026

This comparison changed again in less than a week. Google expanded Lyria 3.5 into Gemini and the Gemini API in early September. Then Suno released V6 on September 9 with a new model family and deeper creation and editing controls. The useful question is no longer “which model sounds better?” It is which environment gives you the kind of control your project actually needs?

Editorial note: This is an evidence-led workflow comparison based on current official product information, not a claim that Jack Righteous has completed an equivalent controlled listening test across every V6 and Lyria 3.5 route. Performance judgments should be added only after like-for-like prompts and revisions are tested.

What Google Lyria 3.5 currently brings

Google describes Lyria 3.5 as an advance in musicality, lyrics, vocals and creative control. It says the model can create richer melodic structures, produce higher-quality lyrics with better prompt adherence and structural awareness, deliver more expressive vocals with improved pronunciation, and give users easier control over tempo and duration.

Google’s current Lyria page describes its most advanced music model as capable of tracks up to about three minutes, while the Gemini API documentation describes full-song generation with structural coherence and high-fidelity 44.1 kHz stereo output from text or image input. Google also offers a separate short-clip route.

The access story is a major strength. Lyria 3.5 is available through the Gemini app, Flow Music, Google AI Studio / Gemini API and Google Vids. That means the same model family can serve casual generation, creator workflows, developer systems and broader media work.

The important distinction is that those surfaces are not interchangeable. Google’s current API guide says Lyria music generation is single-turn and does not support iterative editing or refining a generated clip through multiple prompts in the current API version. Flow Music therefore remains the Google surface to examine when the job is ongoing song development rather than API generation alone.

What Suno V6 changes

Suno introduced V6 on September 9, 2026 as a family rather than one model. V6 is the flagship model for reliable, precise creation. V6 Wild is intended for more varied and exploratory results. V6 Mini is the faster model available to all plans.

Across the V6 family, Suno says the models understand more of the language and building blocks musicians use, including vocals, instrumentation, structure, mood, references and overall feel. All three V6 variants support up to eight minutes per generation.

The bigger control story is editing. Suno says V6 can edit part of an existing song using plain-language direction while preserving the rest, build mashups from multiple sources, sample and isolate material, create from text, audio, images and video, and update a single lyric without rebuilding the entire song. Suno’s help documentation also states that V6 powers the current Remix & Edit tools, including Extend, Cover and Replace Section.

That does not eliminate the need for discipline. More capable editing still works best when you identify the keeper, isolate the actual failure and make the smallest useful change instead of repeatedly regenerating the entire song.

See the Jack Righteous Suno hub for the current workflow system. The URL is retained for continuity even as the content is being migrated to V6.

Artist identity: V6 did not erase the continuity layer

Suno’s recurring-identity tools were introduced before V6, but they remain part of the current ecosystem around it. Voices provides a reusable verified voice workflow where available. Custom Models let eligible creators personalize generation around authorized original music, and Suno’s current V6 documentation explicitly lists Custom Models among the tools powered by V6. My Taste and other personalization tools add another layer of continuity around repeated preferences.

This is different from simply remembering how a user likes to prompt. For creators building a catalogue, the practical question is whether the platform helps preserve a recognizable vocal, musical and production identity across more than one song.

Comparison at a glance

The six tests that define real creative control

A feature list can make almost any generator sound powerful. A professional decision needs a better standard.

1. Brief control

Can you define purpose, audience, emotion, genre, vocal character, structure, duration and exclusions before generation? A prompt is useful only when important instructions survive competing requests.

2. Generation control

Does the result follow the brief, or merely resemble its genre words? Test the intro, section order, lyric subject, lead and backing roles, dynamic build and ending separately.

3. Selection control

Can you compare meaningful alternatives and understand why one candidate is stronger? More generations are not automatically more choice when every result changes unpredictably.

4. Revision control

Can you shorten an intro, replace a weak verse, change one lyric, preserve the hook or strengthen the ending without losing everything else? This is where a generator becomes a production tool.

5. Continuity control

Can you carry a recognizable voice and sound across several songs, an album, a campaign or a client project? Suno currently exposes the clearer public recurring-identity layer. Google’s creator workflows remain worth testing for how consistently they preserve artist identity across projects.

6. Proof and export control

Can you retain the brief, lyrics, prompts, candidates, revision history, exports and applicable terms? A usable track is stronger when the creator can explain how it was directed and developed.

True control is not the number of settings shown before generation. It is how much of your intention survives from the brief to the finished asset.

Where Lyria 3.5 could be the better choice

Lyria’s strongest advantage is the breadth of Google surfaces around it. A creator can start with low-friction generation in Gemini, move into Google’s creator ecosystem, use Lyria inside content workflows such as Vids, or build a custom system through the Gemini API.

For developers, that API access is a meaningful differentiator. But do not mistake API availability for a full iterative editing environment: Google’s current music-generation documentation says multi-turn editing and refinement of generated clips are not supported in the present Lyria 3.5 API workflow.

For creator-side revision, Flow Music remains the Google route that deserves the most direct testing. The important questions are whether edits preserve the rest of the song, whether vocals remain consistent and whether repeated changes introduce artifacts or musical drift.

Where Suno V6 currently has the stronger case

Suno’s advantage is the amount of song-development work that lives inside one music-focused environment. V6 generation, up-to-eight-minute outputs, plain-language local edits, Remix & Edit tools, extensions, section repair, uploaded material and identity-oriented tools all sit inside the same creator ecosystem.

V6 also expands the kinds of instructions Suno says it can understand. The release specifically highlights editing one section while preserving the rest, combining elements from multiple songs, sampling and isolating material, creating from multiple media types and changing a single lyric without rebuilding the full song.

That makes Suno the stronger starting recommendation for creators whose main goal is songwriting, iterative song development, recurring vocals or catalogue continuity—provided those are the things they actually need.

The fair test both platforms should face

Do not compare a carefully developed Suno song with the first Lyria result, or vice versa. Use one project brief, the same creator-written lyrics where the workflow permits, and the same exclusions. Then run four rounds:

1. First generation: judge adherence, not personal excitement.
2. Revision: shorten the intro, remove one unwanted element and strengthen the final section using the best supported revision route on each platform.
3. Continuity: request a second song intended for the same artist or campaign.
4. Documentation: record prompts, settings, candidates, revisions, exports and the terms applying on the creation date.

A useful scorecard should weight brief adherence, lyrics and structure, vocal direction, arrangement, revision control and repeatability. Initial audio quality should not be allowed to hide a weak workflow.

Which route should you use?

Choose Gemini first when:

• you want a low-friction conversational route to a new track;
• you need a quick backing track, jingle, ringtone or concept;
• you want a music result without entering a deeper production environment.

Choose Flow Music first when:

• Google’s creator ecosystem is central to your workflow;
• you want to evaluate Google’s deeper song-development and production experience;
• your project already uses Google’s broader creative tools.

Choose the Gemini API / AI Studio when:

• you are building a custom creator tool or repeatable application workflow;
• music generation needs to happen inside software rather than a manual consumer workflow;
• you understand that current Lyria API generation is single-turn rather than a multi-turn edit session.

Choose Suno V6 first when:

• you are developing complete songs from custom lyrics or detailed direction;
• you need focused revisions, Replace Section, Extend or other V6-powered edit tools;
• longer single-generation song length matters;
• a reusable voice, custom model or recurring artist identity matters;
• you are building a catalogue or album rather than a one-off asset.

Commercial use, copyright and provenance are different questions

A platform may permit a type of commercial use without guaranteeing copyright protection or eliminating third-party claims. Before release, separate these questions:

1. What does the current plan permit?
2. What contractual rights does the platform grant or retain?
3. Which parts of the work reflect protectable human authorship?
4. What provenance or watermarking applies?
5. Did any input involve material you were not authorized to use?

Google DeepMind states that audio generated by its Lyria models is imperceptibly watermarked with SynthID. That is a provenance measure; it does not answer the separate questions of commercial permission, copyright eligibility or infringement risk.

Keep the original brief, creator-written lyrics, prompt history, candidates, selected output, revisions, DAW edits, human recordings and the terms in effect on the generation date. The free AI Song Development Workbook and Bee Righteous Rights + Contribution Tracker are designed to help turn that principle into a repeatable habit.

This section is educational information, not legal advice. Review the current terms that apply to your account and intended use.

Final verdict

Lyria 3.5 matters because Google is competing across access points: Gemini for ordinary users, Flow Music for creator work, the Gemini API / AI Studio for builders and Google Vids for broader media workflows.

Suno V6 answers a different part of the problem especially well: it brings long-form generation, precise and exploratory model choices, focused local editing, Remix & Edit tools and recurring creator-identity systems into one music-first environment.

The winner changes by job. For quick Google-native generation or custom software, Lyria 3.5 has a strong case. For iterative song development and recurring artist identity, Suno V6 currently presents the more complete integrated path. The professional approach is to choose the environment by job and judge what can be directed, revised, repeated and documented—not what produces the flashiest first demo.

Frequently asked questions

Is Lyria 3.5 only available through Flow Music?

No. Google makes Lyria 3.5 available through multiple surfaces including Gemini, Flow Music and developer tools.

Can the Lyria 3.5 API iteratively edit a generated song through repeated prompts?

Not in Google’s current documented API workflow. Google states that Lyria music generation is single-turn and that iterative editing or refinement of a generated clip through multiple prompts is not supported in the current version.

What is the difference between Lyria 3.5 and Flow Music?

Lyria 3.5 is Google’s music-generation model. Flow Music is a creator-facing Google environment around music creation and development.

Can Lyria 3.5 generate lyrics and vocals?

Yes. Google describes improvements in lyric quality, prompt adherence, structural awareness, vocal realism, emotional expression and pronunciation.

What changed in Suno V6?

Suno introduced three V6 variants, up to eight-minute generations and broader V6-powered creation and editing capabilities. Suno specifically highlights focused edits to existing songs, multi-source mashups, sampling/isolation workflows, multimodal creation and single-lyric changes.

Does Suno still support recurring creator identity?

Yes. Suno’s current ecosystem includes reusable voice and personalization workflows, and its V6 documentation explicitly lists Custom Models among the tools powered by V6. Eligibility can vary by plan, feature and location.

Can AI-generated music be used commercially?

Commercial permission depends on the current platform terms, plan and intended use. Permission from a platform is separate from copyright eligibility and third-party infringement risk.

Primary sources reviewed for this September 12 update: Google, Lyria 3.5 in Gemini; Google DeepMind, Lyria overview; Google AI for Developers, Lyria 3.5 music generation; Suno, Introducing V6; and Suno Help, What’s New in V6.