Lyria 3.5 Pro is Google's new preview full-song model with text/image input, 44.1 kHz stereo output, vocals and $0.08-per-song API pricing.
Lyria 3.5 Pro is Google's newest preview music model for full-length songs with coherent verses, choruses, bridges and vocals.
Lyria 3.5 Pro is Google’s flagship Lyria 3.5 music-generation model, released in public preview on September 3, 2026.
Google documents text and image input, 131,072 input tokens, high-fidelity 44.1 kHz stereo MP3 output, timed lyrics and full-song structures with verses, choruses and bridges.
It is available through the Gemini API as lyria-3.5-pro-preview at $0.08 per song.
It is best for full-song demos, vocals, soundtrack concepts and creative composition where structure matters.
It is still a preview model, there is no free API tier, and generated music should be reviewed for quality, policy and rights considerations.
In addition, its main capabilities include Full-song generation, 44.1 kHz stereo, Natural vocals, Timed lyrics, Verses/choruses/bridges and Text and image prompting. For example, common use cases include Music demos, Songwriting, Creative production, Soundtracks and Marketing audio.
In practice, this model may suit Full-song generation, Music demos, Vocals, Lyrics, Creative composition and Image-guided music. Also, notable strengths include Newest Google music model, Full-length structure, High-fidelity stereo and Image-conditioned generation. However, review trade-offs such as Preview endpoints can change, Generated music requires rights/policy review and Exact language coverage is not fully specified before adopting it.
Meanwhile, $0.08 per full-song generation on the paid Gemini Developer API tier. Google charges $0.08 per Lyria 3.5 Pro song generation; there is no free API tier for this preview model.
Use the official model website, official documentation, pricing or release source and additional primary source to confirm current availability, limits and pricing. Product details can change after publication, so rely on primary documentation for final decisions.
Next, continue your research in the AI models directory, Google models and Music Generation Model models. Compare providers, pricing, modalities and practical limitations side by side to choose the right model for your workflow.
First, test the model with a small set of realistic tasks before relying on it for production work. Also, check response quality, consistency, latency, supported file types, context limits and the effort required to review its output. For sensitive or regulated work, examine the provider’s privacy, data-retention, regional-processing and security documentation before submitting private information.
However, AI systems sometimes return incomplete, outdated or confidently incorrect information. Therefore, check important claims against trusted sources and test generated code before deployment. Pricing, quotas and model availability can also change without notice. Finally, revisit the official documentation before you plan a long-term integration or a large-volume workload.
Create a full cinematic pop song with two verses, a chorus and a bridge.Generate a soulful jazz track inspired by this image, with natural vocals.Write and perform an upbeat electronic song with a two-minute structure.$0.08 per full-song generation on the paid Gemini Developer API tier.
A timely addition for the directory: Google’s newest full-song generation model is inexpensive per request and much more structured than clip-focused music models.
EMEleven Music v2 is ElevenLabs’ studio-grade music model for full songs with vocals, multilingual lyrics and section-by-section editing,…
PMPika Music is Pika’s music model that creates finished tracks from text, lyrics, a reference voice or an…
SVSuno v5.5 is Suno’s music model for complete songs with vocals and custom voices and styles, on Pro…
Suno v3.5 is Suno's Summer 2024 music model with improved song structure and up to four-minute initial generations.