Modular MAX Now Supports Apple Silicon GPUs

💻 Modular's MAX platform now supports GPUs in Apple Silicon chips (from M1 to M5).

Users can now run text LLMs (e.g., Qwen 3.5), computer vision models, and generative models (such as FLUX.2 [klein] 4B) directly on Mac GPU cores. Direct GPU support allows for efficient utilization of the unified memory architecture.

🌍 Expanding the high-performance model inference ecosystem to consumer Apple hardware reduces developer reliance on proprietary frameworks and cloud GPUs for local debugging.

👤 It is now possible to run modern, heavy models like FLUX.2 directly on a MacBook using the GPU accelerator, significantly speeding up content generation and local AI workflows.

Source 1: https://forum.modular.com/t/max-models-can-now-run-on-apple-silicon-gpus/3283