MiniMax simultaneously introduced "MiniMax Music 2.0" on social media and its official website, positioning it as an "AI composer, singer, and producer." It emphasizes realistic vocals, cross-genre capabilities, and emotional expression, showcasing features such as duets and a cappella performances. Marketing materials claim it can generate up to 5 minutes of audio, with controllable multi-instrument arrangements and professional-grade sound quality. Meanwhile, the official API documentation lists the model name "music-2.0," providing an interface called "/v1/music_generation," requiring input of style/emotion descriptions and lyrics, and allowing selection of formats such as MP3 and a 44.1kHz sampling rate. Output can be a temporary URL or HEX audio data.
It's important to note that specific metrics such as length limits, fine-grained control over duets and multi-instrument performances are not quantified item by item in the documentation; please refer to the product page and subsequent updates for the most accurate information. Regarding the verifiable aspects, it is currently confirmed that it supports generating complete songs with full vocals and accompaniment directly from text and lyrics, that basic audio parameters are configurable, and that a standard authentication process is available for developers.
Frequently Asked Questions
Q: What are the main capabilities of MiniMax Music 2.0?
A: It is designed for integrated creation of composition, singing and production, covering styles such as pop, jazz, blues, rock, folk, etc., emphasizing realistic singing and emotional control, and supporting a cappella or multi-part singing.
Q: Does it really support 5 minutes and fine control of multiple instruments?
A: The marketing copy contains relevant statements, but the API documentation does not specify the precise upper limit and control granularity. It is recommended to refer to the official product page and subsequent instructions.
Q: How can developers integrate this feature?
A: Call "/v1/music_generation", with parameters including model=music-2.0, prompt (style/scene/mood), lyrics (can include [Verse]/[Chorus] structure tags), and audio_setting (sample rate/bitrate/format). Output supports URL or HEX.
Q: To what extent can the sound quality parameters be configured?
A: The document example shows 44100Hz sampling, 256kbps bitrate, and MP3 format; the linked results have a limited validity period and need to be downloaded and saved in time.
Q: How are compliance and authorization handled?
A: Please refer to MiniMax's official terms and conditions; the copyright, commercial use, and use of materials for generated content must comply with platform policies and relevant laws.
