OpenAI has launched a 28-day improvement initiative aimed at delivering continuous and tangible optimizations for Codex and Work users. If an update is not completed on a given day, user quotas will be reset. On the first day of updates, the default speeds of GPT-6 Astra and GPT-6.1 Sol under the subscription system were increased by approximately 50%. The following day, OpenAI rolled out four significant updates spanning Codex, ChatGPT, and API domains: The Auto-review feature is now freely available to all logged-in ChatGPT users. By introducing a dedicated review Agent, an automated verification layer has been added between users and high-risk operations, reducing the burden of frequent manual approvals during long tasks. API paid usage tiers have been simplified from five levels to three: Build, Launch, and Grow. The cumulative payment threshold for the highest tier has been reduced from $1,000 to $500, with automatic upgrades requiring no application, lowering the barrier for developers transitioning from testing to large-scale production calls. On the macOS version of the ChatGPT desktop app, Pro and Business users can now access the Meetings plugin Beta. This plugin automatically records meeting content and synchronizes outputs to ChatGPT Space, enabling contextual continuity between meeting scenarios and subsequent task execution. Additionally, the Decisions API, previously showcased at DevDay, has opened for public beta testing. Powered by GPT-6 Luna and designed specifically for judgment-based tasks, it operates up to 10 times faster than standard generative APIs, significantly enhancing efficiency for judgment-heavy calls in Agent workflows. These updates do not solely focus on enhancing model capabilities but also aim to address practical operational challenges encountered by Agents in daily work applications.