Google's Lyria 3 and MiniMax Audio's music models are both serious AI music generators, but they come from opposite corners of the industry. One is Google DeepMind's research-grade system riding Gemini; the other is a Chinese AI lab's commercially priced platform with a mature API. Most comparisons stop at "both make songs." This one goes deeper: what each tool is actually built for, where the specs are documented, and which one fits your workflow.
The Lyria facts below come from Google's announcements and the DeepMind model card. The MiniMax facts come from MiniMax's official platform documentation and pricing pages — treat anything version-specific as "verify before you rely," since both sides ship new model versions frequently.
The Short Answer
If you need a precisely structured, high-fidelity song from a detailed text brief with Google-ecosystem integration, Lyria 3 Pro is the stronger fit. If you need a priced, API-first music generation platform with fast, cheap generation and reliable developer docs, MiniMax Audio (music-3.0) is a genuinely competitive pick — especially if you're building a product, not just making tracks.
What Each Tool Actually Is
Lyria 3 / Lyria 3 Pro is Google DeepMind's music model. Base Lyria 3 makes up to 30-second clips; Lyria 3 Pro, announced March 25, 2026, writes full songs up to 3 minutes with intros, verses, choruses, and bridges, AI vocals in 8 languages, and accepts text, lyrics, and up to 10 reference images. No audio upload, no voice cloning, no stems. Every output carries SynthID + C2PA.
MiniMax Audio is the audio product line from MiniMax, a Chinese AI company. It includes TTS (speech models), voice cloning, voice design, and music generation (the music series). Its current music model is music-3.0 (released July 16, 2026), positioned around "intent understood, sound elevated, vocals humanized." Earlier models in the line reached 4-minute songs. MiniMax's music API is metered: $0.15 per up-to-5-minute track with a free tier at 3 RPM, per its official pricing page. MiniMax also offers cover generation (music-cover), style transfer, and auto-lyrics extraction.
The core difference in one line: Lyria is built to execute a precise brief faithfully inside the Google ecosystem; MiniMax is built as a priced, API-first music generation platform.
The Decision Table
| Criteria | Lyria 3 Pro | MiniMax Audio (music-3.0) |
|---|---|---|
| Max track length | Up to 3 minutes (official) | Up to 5 minutes per generation ($0.15/track pricing implies this); earlier models reached 4 minutes (verify) |
| Structure control | Strong — sections + documented timestamp technique | Good — intent-driven; cover/structure tools on music-cover (verify) |
| Vocal quality | AI vocals in 8 languages, high-fidelity | Vocals "humanized" per positioning; strong TTS pedigree (verify) |
| Stems / DAW export | Not available today | Not prominently documented for music models (verify) |
| Input types | Text, lyrics, up to 10 reference images | Text prompts, lyrics, reference audio for covers (verify) |
| Voice cloning / audio upload | No | Yes — MiniMax's TTS line offers 10-second voice cloning (verify) |
| Watermark / provenance | SynthID + C2PA on every output | Verify current policy |
| API pricing | Pay-per-generation on Gemini API / Vertex AI | $0.15 per up-to-5-minute music, free tier 3 RPM (verify) |
| Access model | Gemini app, API, Vertex AI, ProducerAI, credit studios | Web app + API platform |
Two rows deserve extra weight. Input flexibility: MiniMax's platform spans TTS, cloning, and covers alongside generation, so it's a fuller audio toolkit. Price transparency: MiniMax publishes a clear per-track price, while Lyria's cost depends entirely on your access path.
Where Lyria Pulls Ahead
Based on Google's documentation and hands-on use of the Pro model:
- Faithful execution of a detailed brief. Genre, mood, instrumentation, tempo, vocal style — Lyria tends to render what you asked rather than reinterpret it.
- Timestamp-driven arrangement. You can script
[00:00] … [03:00]cues in the prompt so the bridge lands where you want it. Google documents this in its official prompting guide. - Image-guided generation. Up to 10 reference images to steer mood is a genuinely different creative on-ramp.
- 8 documented vocal languages. English, German, Spanish, French, Hindi, Japanese, Korean, Portuguese — explicit and official.
- SynthID + C2PA provenance on every track.
If you want Lyria 3 Pro without a Google Cloud project or a paid Gemini plan, the credit-based Lyria 3 Pro generator on this site runs it pay-as-you-go — and our Lyria 3 Pro guide covers the full capability set.
Where MiniMax Pulls Ahead
MiniMax's strengths cluster around platform breadth and API economics:
- Published per-track pricing. $0.15 per up-to-5-minute track with a free tier is transparent and easy to budget (verify current rates).
- Longer generations. Up to 5 minutes per track at that price point — useful for soundtrack-length pieces.
- Full audio toolkit in one place. TTS, 10-second voice cloning, voice design, covers, and music generation — if your product needs more than music, MiniMax has it.
- Cover generation and style transfer. Reference-audio covers with lyrics modification is a capability Lyria doesn't offer.
- Mature developer docs. MiniMax's API documentation and rate limits are thorough and publicly available.
Choose Lyria 3 (Pro) If…
- You have a specific, detailed brief and want it executed faithfully.
- You need precise song structure — a chorus that hits at a set point, a defined bridge.
- You want to steer with reference images or generate vocals in one of the 8 supported languages.
- You already work inside Google's ecosystem — or want Lyria on pay-as-you-go credits without a subscription.
Choose MiniMax Audio If…
- You're building a product that needs music generation via a priced, documented API.
- You want longer tracks (up to 5 minutes) at a clear per-track price.
- You need covers, style transfer, or voice cloning alongside generation.
- You prefer a single vendor for TTS, cloning, and music.
Rule of Thumb
Match the tool to the delivery pipeline, not to a single demo track. If the deliverable is a structured song from a written brief at high fidelity, lean Lyria 3 Pro. If you're embedding generation in a product with per-track cost targets, lean MiniMax.
How to Tell You Chose Right (A Low-Friction Test)
Run the same detailed prompt through both once and score two things:
- Faithfulness: how many of your explicit requests (tempo, instrument, bridge position) made it into the output?
- Cost & integration fit: can you get the format, license, and API economics you need for where this ships?
If the track that sounds best doesn't fit your cost or workflow, it's not the winner for your project.
FAQ
Is Lyria 3 better than MiniMax for music? Not universally. Lyria 3 Pro leads on faithful brief execution, timestamped structure, image-guided generation, and 8 documented vocal languages. MiniMax music-3.0 leads on published per-track pricing, longer generations, and a broader audio toolkit (covers, cloning, TTS).
Lyria 3 vs MiniMax for API use — which wins? MiniMax has a transparent $0.15 per up-to-5-minute music API with thorough developer docs. Lyria is available via the Gemini API and Vertex AI with pay-per-generation pricing. If you need predictable per-track cost and a single audio vendor, MiniMax is easier to budget; if you're already in Google Cloud, Lyria integrates naturally.
Can I use tracks commercially? Lyria's commercial rights depend on the surface you generate on — check the terms of that path. MiniMax's commercial terms should be checked on its official site and docs as well; both require reading your specific plan's terms.
Which has better vocals? Lyria 3 Pro documents 8 vocal languages with high-fidelity output. MiniMax's speech line is its best-known strength (10-second cloning, 40-language TTS), and music-3.0 markets "humanized" vocals — but its music vocal spec is less explicitly documented. Test with your own material.
What's the price difference? Lyria is free in the Gemini app, metered on API/Vertex, and credit-based on studios. MiniMax music is $0.15 per up-to-5-minute track with a free tier at 3 RPM (verify current rates). For heavy API volume, MiniMax's per-track price is easy to calculate.
The Bottom Line
Lyria 3 Pro is the better brief-executor: precise structure, documented fidelity, image inputs, and multilingual vocals inside Google's ecosystem. MiniMax music-3.0 is the stronger API platform: transparent per-track pricing, longer generations, and a full audio toolkit including covers and cloning. Decide by your pipeline, not by a single listen. If Lyria is your pick and you'd rather skip Google Cloud setup, run the Pro model on credits through the Lyria 3 Pro generator here — and start with our Lyria 3 overview for the full model breakdown.
Sources
- Google — Lyria 3 Pro announcement: https://blog.google/innovation-and-ai/technology/ai/lyria-3-pro/
- DeepMind — Lyria model page: https://deepmind.google/models/lyria/
- DeepMind — Lyria 3 model card: https://deepmind.google/models/model-cards/lyria-3/
- MiniMax — Models overview (music-3.0): https://platform.minimax.io/docs/guides/models-intro
- MiniMax — Pay-as-you-go pricing: https://platform.minimax.io/docs/guides/pricing-paygo
Update note: MiniMax ships frequent model versions and pricing changes — the figures above are from MiniMax's official docs as of August 2026 and must be verified on their platform pages before relying on them.
