Last released May 28, 2026
Ultra-fast LLM inference engine with a Vulkan compute backend
Supported by