When the Model Gets Cheaper, the Risk Surface Gets Wider

GLM 5.2’s rise against Claude Opus is more than a benchmark story: it shows how open-weight AI is changing the economics of agentic systems, where task completion cost matters more than token price alone.