Announcement

Photon compiles models into GPU programs optimized for the chip and the job. The first release supports Moondream, Qwen, and Gemma on NVIDIA H100, and wins every matched throughput test against vLLM and SGLang.

Photon 1.3.0 makes Moondream faster across NVIDIA, Mac, and Windows, runs finetunes on far more hardware, fixes an accuracy issue on older GPUs — and running Moondream locally is now completely free.

Photon 1.2.0 brings native inference to Apple Silicon and Windows, adds NVIDIA Blackwell and Jetson Thor support, and ships meaningful speed gains across existing GPUs.

Solve the last-mile problem with Lens, our fine-tuning product that makes VLMs production-ready.

Photon brings real-time Moondream inference to production vision AI, from edge devices to H100-class servers.

Mac users can now run Moondream 3 Preview in Station with MLX-native, quantized performance.

Fast, cheap, smart. Pick three.

Moondream now supports 4-bit quantization, making it faster and smaller.

Moondream Station is a one-click solution to running Moondream locally.