The Frontier AI Inference Cloud for Agents — Byung-Gon (Gon) Chun, FriendliAI
Open-weight models have reached frontier quality and run agentic workloads ~5.6x cheaper than closed models.
“Open weight models have crossed the quality threshold”
FriendliAI CEO Byung-Gon Chun argues 2026 is the year agents go into massive production, driven by exploding agent adoption and open-weight models reaching frontier capability. He claims open models like GLM 5.2 now match closed models such as Opus 4.8 on real agentic tasks while costing a fraction (27 cents vs ~$150), and that the inference stack must be rebuilt for agent workloads rather than chat.