Skywork AI launches the world's first music reasoning large model Mureka O1, leading the global AI music revolution

SINGAPORE, March 26, 2025 /PRNewswire/ -- On March 26th,  Skywork AI officially launches the Mureka O1 and Mureka V6 models. As the world's first music reasoning large model, Mureka O1 outperforms Suno and achieves the State of the Art (SOTA). 

In April 2024,  Mureka released the first-generation music model Mureka V1 (SkyMusic). It received an overwhelmingly positive market response.

Over the past year, our technology and product teams have been working hard to push the boundaries and refine Mureka's capabilities. So far, users from more than 100 countries and regions around the world have accessed our AI music product—Mureka.

Mureka is a combination of "Music" and "Eureka." Eureka is an exclamation from Greek, meaning "I have found it" or "I have discovered it," often used to express moments of sudden inspiration. Mureka hopes that it can quickly capture people's moments of musical inspiration and create songs in their own unique style.

Mureka V6 & O1 have been fully launched and support seamless switching

Mureka V6 is the current base model of Mureka, supporting pure music generation as well as AI music creation in ten languages, including English, Japanese, Korean, French, Spanish, Portuguese, German, Italian, Russian and Chinese. In Mureka V6, we have introduced our self-developed ICL (In-Context Learning) technology. This advancement brings several improvements: a more expansive soundscape, enhanced vocal texture, and improved mixing design.

The Mureka O1 model is an optimized version based on Mureka V6 and is the world's first music model to introduce CoT (chain of thought). It incorporates thinking and self-criticism in the inference process, significantly improving music quality, music creation efficiency, and flexibility.

Both Mureka V6 and O1 models support a wide range of music styles and emotional expressions, which cover Jazz, Electronic, Pop, Country, R&B, Soul, Blues, Rock, Dance, and more. The emotional dimensions include happiness, indulgence, mystery, vitality, sadness, and various other emotional expressions.

In addition, Mureka also offers two unique music generation features:

1. Music Reference Function: Using the music itself as a prompt, users can directly upload audio files or YouTube links as creative references.
2. Vocal Function: Mureka is the world's first AI music generation platform that allows users to choose the vocal timbre of the singer. Users can not only choose from a variety of official singer timbres but also upload their own voices for the AI to learn and replicate, accurately simulating the singer's timbre and generating personalized exclusive works with a single click. The feature of customizing singer timbres heralds the arrival of the era where everyone can become an AI singer.

The launch of Mureka V6 and O1 represents a significant milestone in the technological advancement of the AI music industry. It also demonstrates the potential of AI in artistic creation and highlights the vast possibilities of continuously exploring new technological forms and artistic expressions.

Mureka O1 adopts the self-developed MusiCoT technology and achieves SOTA

Mureka O1 includes MusiCoT, Mureka's latest innovative research in the field of music generation, which is technologically advanced and forward-looking. We found that there is very little algorithmic work in music, and only a few teams have made their work public. To address this,Mureka published the implementation principle of Mureka O1 in the form of a paper, and welcome the industry to make better work on this basis.

Mureka O1 incorporates the latest achievements in music generation from the Mureka team—MusiCoT. MusiCoT is a variant of the Chain-of-Thought (CoT) method tailored for music generation. Unlike traditional autoregressive models that generate audio step-by-step, MusiCoT pre-generates the overall music structure before finalizing the audio tokens for decoding. This approach significantly improves the structural coherence and instrumental arrangement of the generated music.

Built on the CLAP model, MusiCoT achieves high scalability without requiring manual annotations. It not only enhances the analyzability and quality of generated music, but also paves a new path for high-fidelity music generation. With MusiCoT, we usher in an era of structured music creation using AI, offering exceptional possibilities for musicians and producers.

Given to the introduction of Chain-of-Thought (CoT) technology and the upgrade of the algorithm framework, Mureka O1 not only maintains low-latency music generation but also significantly improves the lyrics accuracy, singing accuracy, and production quality. It surpassed Suno V4 in multiple metrics.

Music isn't math. Ultimately, it's all about what sounds good to the ear. That's why both subjective evaluation and objective metrics are crucial. Under this philosophy, Mureka O1 excels across various aspects in our subjective assessment and ranks among the top-tier models in this category，including Mixing, Vocal Texture and Background Music Texture.

Mureka offers API service and model fine-tuning function

Mureka is the world's first high-quality AI music generation platform with open API services. Both developers and music platforms can now seamlessly integrate Mureka's music generation capabilities into their own products or platforms, making it easier to extend their business value with AI music capabilities.

In addition to opening up API services for the first time worldwide, Mureka synchronously opens up model fine-tuning services based on Mureka V6. Become the world's first AI music generation platform with open model fine-tuning function.

This feature makes AI music creation more personalized, providing customized AI music solutions for musicians, producers, and even brands and game developers.In the future, the Mureka team will continue to invest in model capabilities to keep our music model in the top tier in the world, while developing more music creation functions on top of the base model, such as secondary editing of music generation, song creation, and so on.

Users around the world are welcome to log on to Mureka (www.mureka.ai) to experience the new V6&O1 model, explore the infinite possibilities of music creation with AI music creators around the world, and build an AI music developer ecosystem.

SOURCE Skywork AI