Codex 改进挑战第 12 天：Astra Pro 提速与并行 agent
用户吐槽 Day 12 更新只讲缓存与路由优化却不给额度数字，同时提到 GPT-6 Astra Pro 响应更快、工具调用更强，Pro 用户可在 ChatGPT 并行跑多个 agent。
🚨 Day 12 of OpenAI's Codex improvement challenge is UNDERWHELMING for users. Two things announced today: 1. Codex and Work get more effective usage through better caching, context reuse, and request routing. But OpenAI gave us NO NUMBERS on how much extra usage we're getting. 2. GPT-6 Astra Pro gets faster responses and better tool use. Pro users can also run multiple agents in parallel in ChatGPT. Here's what pisses me off: Better efficiency sounds good, but we still don't know how much it actually improves our usage. I don't expect OpenAI to ship a groundbreaking feature every day. BUT if you're promising daily improvements, show us the impact. Some progress, but usage transparency still gets zero points. Day 12: 50 points TIBO, Give us Resets as Promised. 😭