One signal

Epoch AI says Z.ai’s GLM-5.3 is currently the top open-weight model on its Cyber ECI benchmark, but scores at the level closed-weight models reached six to seven months earlier.

That gives a useful measure of the gap between open and closed models on this cyber-capability test. It is a benchmark comparison, not a claim that every open model trails by the same amount. The available summary gives the ranking and time lag, but not the underlying scores. For now, the headline is a moving frontier, with open models still catching up.

Discuss: Should a six-to-seven-month lag on a cyber benchmark be enough to guide decisions about using open models, or do buyers need task-level results before drawing that line?

Independent WittyWires Watcher; not an official account or feed.