Codex 0.156.1

當你連續請 AI 撰寫程式碼而發出大量請求時，最怕遇到對話中斷而打亂思緒。在模型選單（挑選由哪一款 AI 回答問題的切換清單）中，現在可以直接指定 GPT-6 Sol 或 GPT-6 Luna。當發送的請求達到上限並觸發限流（短時間內發送過於頻繁時暫停接收的保護機制）時，畫面跳出的切換提示現在會主動推薦改用 GPT-6 Luna 接手。

如此一來，在密集調度不同模型資源時，便能直接順著建議切換，不必中斷手邊正在進行的開發任務。

When sending numerous requests while asking AI to write code consecutively, unexpected interruptions that break your train of thought are a major hassle. From the model picker, you can now directly choose GPT-6 Sol or GPT-6 Luna. When outgoing requests hit their cap and trigger rate limits (a protective throttle that pauses incoming requests when sent too frequently in a short period), the pop-up switch prompt now proactively suggests handing off to GPT-6 Luna.

This ensures that when juggling model resources heavily, you can follow the suggestion to switch directly without interrupting your ongoing development work.