Freedom251 › AI Music Generators
MiniMax Music 3.0
What MiniMax Music 3.0 is
MiniMax Music 3.0 is an open-weights model that generates complete songs from a creative concept and optional lyrics in one generation. It supports songs up to five minutes and accepts direction about genre, tempo, key, emotion, vocal delivery, instrumentation, arrangement, and production profile. A template-based prompt enhancer expands short descriptions into fuller instructions. Structured Captions add controls for time signature, use case, emotional contour, groove, changing instrumentation, vocal timbre, harmonies, and effects such as delay or Auto-Tune. Lyrics can use section tags including verse, chorus, bridge, and outro. Output is 32 kHz, 16-bit stereo WAV with vocal generation. MiniMax publishes weights and local instructions for SGLang, Diffusers, and ComfyUI, while a web version is listed. Local inference requires CUDA; CPU offloading can reduce memory needs to about 22 GB, or fit 8 GB cards more slowly. Generation is non-streaming, with stated text and acoustic-frame limits. Commercial use requires the Community License terms.
Who it suits
This model suits creators and developers generating structured songs through the web or local open-weight tools. It is also relevant to commercial users able to follow the attribution and authorization requirements.
What it does well
- Generates complete songs up to five minutes
- Accepts detailed musical and vocal controls
- Outputs stereo WAV audio
- Provides downloadable weights for local use
- Supports commercial use under its license
What to know first
- Local inference requires CUDA
- Generation is non-streaming
- Requests may not exactly match musical results
- Commercial use requires specified attribution
Verdict
MiniMax Music 3.0 offers broad song, lyric, and production controls with web access and downloadable weights. Check the hardware, generation limits, and Community License obligations before local or commercial use.
Facts from MiniMax Music 3.0's own pages
- Audio export formats
- wav · minimax.io, 22 Sept 2026
- Audio output
- The model produces 32 kHz, 16-bit stereo WAV audio. · huggingface.co, 12 Oct 2026
- Availability
- MiniMax says its music models stopped being available through Token Plan and M Plan on August 20, 2026, and directs users to MiniMax Audio or the open-source model weights. · platform.minimax.io, 12 Oct 2026
- Commercial use
- Yes · minimax.io, 22 Sept 2026
- Company
- MiniMax describes itself as a global AI foundation model company founded in early 2022. · minimax.io, 12 Oct 2026
- Company products
- MiniMax says its models power products including MiniMax Code, MiniMax Design, MiniMax Audio, Talkie, and its enterprise and developer-facing Open API Platform. · minimax.io, 12 Oct 2026
- Control caveat
- Section tags and music descriptions guide generation but do not guarantee that tempo, key, instruments, lyrics, or structure will exactly match the request. · huggingface.co, 12 Oct 2026
- Founded
- 2022 · minimax.io, 28 Sept 2026
- Free plan
- Yes · minimax.io, 22 Sept 2026
- Generation limits
- The model card lists a 5,000-token text prompt limit, a 9,000-acoustic-frame audio limit, and non-streaming generation only. · huggingface.co, 12 Oct 2026
- Hardware requirement
- Inference requires CUDA; the model card says automatic CPU offloading can reduce GPU memory needs to about 22 GB, with slower layer-by-layer offloading fitting 8 GB cards. · huggingface.co, 12 Oct 2026
- License
- The MiniMax-Music3 Community License requires commercial products using the software to prominently display “MiniMax-Music3” and requires prior written authorization when aggregate yearly revenue exceeds US$20 million. · huggingface.co, 12 Oct 2026
- Local use
- The official model repository documents downloading the model and running it with Diffusers, including a device-map example using CUDA. · huggingface.co, 12 Oct 2026
- Model access
- MiniMax publishes downloadable model weights and instructions for local use with SGLang, Diffusers, and ComfyUI. · huggingface.co, 12 Oct 2026
- Model components
- The model uses an 8B Global LLM for long-range structure, a 0.6B Local LLM for acoustic detail, and a synthesis stack with flow matching and Flow-VAE. · minimax.io, 12 Oct 2026
- Music controls
- Its inputs can specify lyrics and musical details such as genre, tempo, key, emotion, vocal delivery, instrumentation, arrangement, and production profile. · huggingface.co, 12 Oct 2026
- Open weights
- MiniMax describes Music 3.0 as an open-weights model, and its official Hugging Face repository provides model files for download. · huggingface.co, 12 Oct 2026
- Supported platforms
- web · minimax.io, 22 Sept 2026
- Prompt enhancement
- A template-based Prompt Enhancement System can expand a simple description into detailed musical instructions. · minimax.io, 12 Oct 2026
- Purpose
- MiniMax Music 3 generates complete songs from a creative concept and optional lyrics in a single generation. · minimax.io, 12 Oct 2026
- Song length
- The model is designed to sustain creative intent across songs of up to five minutes. · minimax.io, 12 Oct 2026
- Song structure
- Lyrics can use section tags such as intro, verse, pre-chorus, chorus, bridge, instrumental, solo, and outro to define macrostructure. · minimax.io, 12 Oct 2026
- Structured captions
- Structured Captions can specify genre, tempo, time signature, key, use case, production character, emotional contour, instrumentation changes, groove, and vocal delivery. · minimax.io, 12 Oct 2026
- Support
- The Music 3 model card lists [email protected] as its contact address. · huggingface.co, 12 Oct 2026
Where MiniMax Music 3.0 stands on Freedom251
- AI Music Generators#95 of 123