Daily AI Catchup
ModelOpenaiGpt-6-2-AstraCodexChatgpt

OpenAI's next model, GPT-6.2 Astra, spotted in the wild

GPT-6.2-Astra has started appearing in Codex usage breakdowns, likely as part of an A/B test or staged rollout ahead of a public launch that could come within days. OpenAI reportedly skipped a GPT-6.1 Astra release over concerns about deception and scope. Separately, OpenAI is merging ChatGPT's Chat and Work tabs so users can move between casual exploration and task completion in one interface, while Codex stays separate and both will draw from a shared usage allowance.

Why it matters

๐Ÿ’ป Developer ยท Codex users may already be getting routed to a new model without notice. Expect behavior changes in coding tasks before any formal announcement.

๐Ÿ“ฆ Product ยท The Chat/Work tab merger simplifies ChatGPT's product surface and suggests OpenAI is consolidating toward a single unified workspace model.

๐ŸŽจ Design ยท Merging two distinct modes into one interface is a meaningful IA decision worth studying if you design multi-mode AI tools.

๐Ÿ“ˆ Business ยท A shared usage pool across ChatGPT and Codex changes how teams should budget API and subscription spend going forward.

๐Ÿค” Just Curious ยท OpenAI held back a model over deception concerns, then is now quietly testing its successor, a reminder that releases aren't always clean or linear.

Sources: GPT-6.2 Astra appears in Codex usage ahead of possible launch, OpenAI is finally merging ChatGPT's Chat and Work tabs