Meta and Apple Partner to Bring Muse Spark Open AI Model to Mac Studio with 1TB RAM Support

Meta Releases Most Capable US Open-Weight AI Model

Meta and Apple announced a collaboration to launch Muse Spark Open, which Meta positions as the most capable US open-weight model to date1

1

2

2

. The natively multimodal model features 816 billion total parameters with 42 billion active parameters and a 1 million-token context window4

4

. Built specifically for coding and agentic work, Muse Spark Open excels at long-horizon tasks, complex software engineering, and multi-step reasoning1

1

. Meta's benchmarks suggest the model outperforms leading Chinese models GLM 5.3 and Kimi K3 on agentic tasks and visual reasoning2

2

, positioning it as a strategic response to concerns about Chinese AI leadership.

Optimized for Apple Mac Studio with M5 Ultra

Muse Spark Open represents the first model in its class optimized to run locally on Apple Mac Studio machines powered by the M5 Ultra chip1

1

. The model leverages Apple's latest MLX framework, an array framework designed for machine learning tasks on Apple Silicon4

4

. Apple's unified memory architecture allows the M5 Ultra's integrated GPU to use up to 512GB per Mac Studio, meaning two machines can hold an entire model4

4

. With the chip's 1.2TB/s of memory bandwidth, two Mac Studios effectively behave like one supercomputer with 1TB of RAM5

5

. However, running Muse Spark Open locally requires a substantial investment—four 256GB M5 Ultra Mac Studios would cost roughly $38,0002

2

.

Unlikely Alliance Between Former Rivals

The collaboration marks a significant shift in the relationship between Meta and Apple, companies that have clashed for years3

3

. Mark Zuckerberg publicly acknowledged the partnership, stating that Muse Spark Open "runs great on new Mac Studios thanks to work with @johnternus and Apple"1

1

. John Ternus, Apple's new CEO, responded enthusiastically, calling it "amazing to run a model of Muse Spark Open's caliber and scale this efficiently"1

1

. Meta chief AI officer Alexandr Wang told Axios the companies have been collaborating over several months, joking "we're excited for mommy and daddy to be playing nice again"1

1

5

5

. The partnership is particularly notable given their history—Apple's 2021 privacy changes limited digital ad targeting capabilities, with Meta estimating the change would cut its revenue by $10 billion in 20223

3

.

Local AI Model Execution Gains Momentum

The collaboration reflects growing interest in running powerful AI models on local hardware rather than relying on cloud data centers3

3

. Running Muse Spark Open locally on Mac computers adds a layer of security and saves users money by eliminating cloud computing costs3

3

. Meta tested various agentic workflows running entirely on-device using Muse Code powered by Muse Spark Open, including shipping real iOS changes to Ice Cubes, an open-source Mastodon client, with bug fixes and new features4

4

. The model also drove DaVinci Resolve end-to-end, reviewing footage frame by frame to identify strong moments and crop them vertically4

4

. For larger organizations burning through token budgets, investing in Mac Studios to run on Mac computers might prove more economical than cloud-based solutions2

2

.

Availability and Strategic Implications

Muse Spark Open is available now through Hugging Face for download2

2

4

4

. The model is also accessible through API platforms including Baseten, Fireworks AI, Groq, Modal, Nebius, OpenRouter, and Together AI, with weights downloadable from Amazon SageMaker AI, Google Cloud Model Garden, and Microsoft Foundry Models5

5

. As an open-weight AI model, developers can download the technology's parameters from the internet and run the model on various hardware configurations3

3

. Apple published new technical documentation about its MLX framework's ability to split open-source models across multiple machines and revealed new benchmarks for its chips3

3

. The partnership could challenge OpenAI and Anthropic while signaling Apple's growing strength in AI chips, a market Nvidia has dominated due to prevalence in large data centers3

3

. Watch for whether this collaboration extends beyond Muse Spark Open, potentially reshaping competitive dynamics in open-source AI development and on-device AI capabilities.