HeadsUpAI

Cloudflare Adds MiniMax Music 2.6 to Workers AI With Free Inference Week

Cloudflare added minimax/music-2.6 to its Workers AI catalog, a music generation model that creates full-length tracks with vocals from text prompts. The integration follows the MiniMax Music 2.6 launch by bringing high-fidelity music generation to the edge within the Workers environment.
Model
MiniMax Music-2.6
Audio Quality
Up to 44.1kHz
Formats
MP3, WAV
Controls
BPM, Key, Lyrics Optimizer
Availability
Cloudflare Workers AI
Promo
Free inference for one week

This integration provides a programmable alternative to standalone tools, positioning Cloudflare as a competitor to Google's Lyria 3 models. By hosting the model on its global network, Cloudflare reduces the infrastructure overhead required for audio inference. It also builds on its previous Kimi K2.5 integration.

You can generate songs with custom lyrics or use an automated optimizer to write them. The API supports technical parameters including BPM, musical key, and output formats like 44.1kHz WAV. While pricing is usually usage-based, Cloudflare is offering free inference on the model for one week to encourage building audio-first applications.

View the full update on developers.cloudflare.com
Cloudflare
Cloudflare
@Cloudflare
X

Music-2.6 from @MiniMax_AI is free on Cloudflare this week! Generate full-length songs or instrumentals from a text prompt, with optional auto-generated lyrics. Fast global inference on Cloudflare’s network. Perfect for building music apps on #CloudflareWorkers. Try it free before the promo ends: https://t.co/W4uCUjC8vm

5retweets46likes
View on X

Still wondering? A few quick answers below.

MiniMax Music-2.6 is a music generation model available through Cloudflare Workers AI. It allows developers to create full-length songs with vocals or instrumental tracks using text prompts. The model supports technical controls like audio format selection, sample rate adjustments, and specific musical parameters such as BPM and key for precise creative output.

Developers can invoke the model using the standard Cloudflare Workers AI environment. By calling the minimax/music-2.6 model via the AI.run method, you can provide a text prompt describing the style and mood. The API also accepts custom lyrics or can automatically generate them if the lyrics optimizer parameter is enabled in the request.

The model generates high-quality audio in either MP3 or WAV formats. It supports sample rates up to 44,100 Hz and allows for bitrate customization. Developers can also toggle between vocal tracks and pure instrumentals, making it suitable for a wide range of applications from background music to full song production.

While Cloudflare typically charges for Workers AI inference based on usage, MiniMax Music-2.6 is available for free during its launch week starting April 27, 2026. After this promotional period ends, pricing will revert to standard rates which can be viewed directly within the Cloudflare dashboard under the AI models section.

Yes, the model includes a lyrics optimizer feature that can automatically generate song lyrics based on the provided music style and mood description. Alternatively, users can provide their own custom lyrics up to 3,500 characters in length, using newline characters to separate the lines for the vocal generation process.

Share this update