FreeToken: Edge-Native Engine Runs 290B MoE Models on Consumer Hardware

UC Berkeley releases FreeToken, an edge-native MoE serving engine that uses bandwidth-adaptive CPU-GPU co-execution to run 290B-parameter open-weight models on consumer hardware, claiming significantly faster inference than llama.cpp, ktransformers, Ollama and moe-infinity.

2026-08-18 ~ 2026-08-19 · 4 related posts

• FreeToken Claims Faster MoE Inference vs. llama.cpp and Ollama — wavefnx · 2026-08-18
• FreeToken framework claims major MoE inference speedups — wavefnx · 2026-08-18
• UC Berkeley Releases FreeToken for Efficient Edge-Native MoE Serving — UCBerkeley · 2026-08-19
• FreeToken Engine Runs 290B+ MoE Models on Consumer Hardware — wavefnx · 2026-08-19