OpenAI Codex 团队透露产品路线图：Linux 桌面应用、Agent 持久性改进

OpenAI’s Codex team just revealed where the product is heading next.

tl;dr their reddit AMA:

Codex now has over 5 million weekly users - twice as many as three months ago - and shipped 150 improvements during that period.

From the AMA: -A Linux desktop app is in development, but there is no timeline.

-Better agent persistence and lower code complexity are planned for future releases.

-Automatic model routing does not exist today. The preferred UX is Codex inferring task difficulty while preserving a manual speed/reasoning override.

-OpenAI sees Slack, GitHub and Notion connectors as a “step function change” toward making Codex a productive coworker.

-GPT-5.6 was specifically trained to improve UI and frontend work.

-OpenAI made no promise on a 1M-token context window.

-Current pricing is not guaranteed to remain unchanged, though the team says broad accessibility remains the goal. N -ormal ChatGPT conversations do not consume the agentic allowance; Codex and ChatGPT Work do, with costs varying by task.

-OpenAI acknowledged benchmark cheating as a real concern and says it penalizes this behavior during evals while using independent vendors.

you r welcome