AI 导读
现在还早，但 Opus 5.5 medium 目前在 circuit_eval 上表现相当亮眼。Opus 5 在第二个 checkpoint 之后就崩了 每 10 分钟更新一次，直到我的航班起飞 https://t.co/HWfcMjBYP4
正文
its early but opus 5.5 medium is kinda smoking circuit_eval so far. opus 5 blundered after the second checkpoint
updating every 10 minutes until my flight leaves https://t.co/HWfcMjBYP4
来源：@dexhorthy · x.com