01
What is Lyria 3.5?
Lyria 3.5 is Google's flagship music-generation model for creating full-length songs from text prompts or image inputs. Google documents 44.1 kHz stereo output, MP3 audio and text lyrics, with structural control for verses, choruses and bridges.
Google announced broader availability in the Gemini app on September 4, 2026, following the model's public-preview release in the Gemini API on September 3. This is a wider access event, not the original introduction of the Lyria family.
02
Lyria 3.5 in the Gemini app
The Gemini app now lets users create music by describing a genre, choosing vocal or instrumental styles and using creative templates. Google says the experience is available to users globally on web and mobile.
For creators, that makes Lyria useful for rapid soundtrack exploration, backing tracks, jingles, custom themes and other concept-stage audio. The app is the simpler no-code surface, while the Gemini API is intended for programmatic workflows.
03
Lyria 3.5 API access
Google's current developer documentation uses the model ID lyria-3.5 for full-song generation through the Gemini API. The model accepts text and image inputs and returns generated audio plus lyrics text where available.
The model is specialized for music generation. Google's model page marks function calling, search grounding, code execution, structured outputs, Live API and several other general Gemini capabilities as unsupported.
04
Current Lyria 3.5 pricing
Google's Gemini Developer API pricing page currently lists Lyria 3.5 full-song generation at $0.08 per song, with no free API tier for this model.
05
Inputs, outputs and song structure
Lyria 3.5 supports both text prompts and image inputs. That allows workflows where a campaign image, product visual or mood board influences the generated music alongside an explicit written brief.
Google documents full-length songs that can run for several minutes, with duration influenced through prompting. The model is designed for coherent musical structure rather than a fixed one-size-fits-all song length.
06
Creator and marketing use cases
- background music for reels and short-form video
- campaign-specific jingles and intro themes
- prototype audio for ads and client pitches
- music concepts generated from campaign imagery
- rapid soundtrack exploration before final production
These are workflow ideas, not a blanket statement about commercial rights. Paid API access does not by itself settle copyright, ownership, licensing or platform-disclosure obligations.
07
Important limitations
Lyria 3.5 remains a public-preview model as of Google's September 2026 release notes. Preview availability, quotas and behavior can change, so production systems should recheck the live model documentation and project limits before launch.
Google's quality claims should be treated as vendor claims. Independent evidence is not yet strong enough to declare Lyria 3.5 universally better than competing music generators such as Suno or Udio.
08
Where Lyria 3.5 is available
Google lists Lyria 3.5 in the Gemini app and Gemini API, with access also referenced through Google AI Studio and other Google creative surfaces. Consumer-app availability and developer-project access are different surfaces and should not be assumed to have identical account or billing requirements.
For API deployments, verify the active model ID, project billing, quotas and regional/account access immediately before shipping.
Sources
Primary and supporting sources
Facts were rechecked against the linked sources immediately before publication. Pricing, product availability and rollout status can change.