FastFlowLM Joins AMD to Advance AI Inference

Recently, the FastFlowLM team joined AMD, marking another key step in our strategy to advance AI performance and efficiency across the stack. FastFlowLM developed a lightweight, highly optimized inference software flow that delivers fast, efficient large language and multimodal model performance directly on AMD technology-powered AI PCs and workstations.

FastFlowLM was built in the open-source ecosystem and made possible with IRON, the open-source NPU compiler technology developed and released by the AMD Research and Advanced Development (RAD) Group. Incubated by RAD and shaped through engagements with external researchers and developers, IRON underpins a fully open stack for AMD leadership agentic AI platforms.

That open foundation earned rapid traction with a vibrant community of developers and independent software vendors. FastFlowLM’s close alignment and deployment with Lemonade, the AMD open-source inference initiative, has been central to ecosystem adoption, making it easy to bring agentic, retrieval-augmented coding and multimodal experiences to AMD platforms.

The FastFlowLM team joins the AMD Artificial Intelligence Group, where their expertise will accelerate our client and workstation AI software stack and our Day-0 enablement of the latest models. We remain committed to investing in this open ecosystem, and we’re thrilled to build the future of on-device AI together.

More: To explore the latest news and capabilities this team is building, download their recent release of Qwen3.6-35B-A3B – the second mixture of experts model released on AMD NPUs. Learn more on GitHub.