Lyria 3 Pro Preview
Neural Network
Lyria 3 Pro Preview is a Google music generation model that creates full-format tracks with 48 kHz audio via the Lyria 3 API.
Specifications
- Input
- text, images, documents
- Output
- text
- Context size
- 1 048 576 tokens
- Max answer length
- 65 536 tokens
- Capabilities
- Batch API
How it works Lyria 3 Pro Preview?
Lyria 3 Pro Preview is part of the Google Lyria 3 music model family, available via the Gemini API. The model creates music from text descriptions and outputs full-format compositions with 48 kHz audio, meaning the result is ready for mixing and publishing, not just as a rough demo. In BotHub, you can use Lyria 3 without a VPN or foreign card: payments are made with Russian cards in rubles and charged per generation, with unused funds not expiring. Over 250 models are available in one window, so you can create a video script, cover art, and track in one go, with requests sent via a unified OpenAI-compatible API with AES-GCM encryption. Teams have access to contracts, invoices, electronic document management, and an admin panel with limits. It is useful if you are preparing background music for videos and podcasts, assembling a soundtrack for a game or app, making jingles and audio ads, testing musical ideas before studio recording, or iterating on arrangements and moods for client presentations.Frequently asked questions about Lyria 3 Pro Preview
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
lyria-3-pro-preview by Google is a model that accepts text, images, and documents as input. It is suitable for analyzing large materials, extracting data from scans and tables, preparing summaries, and batch processing tasks. In BotHub, it is accessible from Russia: no VPN or foreign card required.
In a single response, the model outputs up to 65,536 tokens, and the same volume is available in the context. This is enough to get a long coherent text, report, or analysis of a large document in its entirety without manually breaking it into parts.
There is no indication of step-by-step reasoning before answering in our data, so we will neither promise nor deny this capability. If you specifically need a step-by-step task analysis, ask for it directly in your prompt to the model.
There is no specific note about function calling and structured JSON output in our data, so we cannot confirm this. If your project requires this specific mode, check out other models in BotHub with a unified API where these capabilities are explicitly stated.
Yes, in addition to text, the model accepts images and documents: you can pass them along with your request to get a description, content analysis, or summary. Audio and video input is not noted in our data, so we will not promise such uploads.
We do not have separate quality measurements for the Russian language for this model, and we will not invent them. However, BotHub itself works from Russia: access without VPN or foreign card, payment in rubles, and you can switch to another model right on the page.