Tom's Hardware reports that Moonshot AI's Kimi K3 debuted at number one on the Frontend Code Arena at 1679 Elo, ahead of Claude Fable 5 — a large jump from where its predecessor sat. The model is a 2.8-trillion-parameter mixture-of-experts system with a one-million-token context window and native multimodal input, making it the largest open-weight model shipped to date. CNBC frames it as the latest sign that Chinese labs are closing the gap with US frontier labs while working around compute limits. The API is live now; Moonshot has dated the full open weights for July 27. Caveat worth keeping: broader benchmark aggregates still place K3 behind the top closed models, so this is a category win, not an outright frontier crown.