In this update cycle, ChatGPT is focusing on “stronger model selection” and “more natural voice interaction.” If you frequently use ChatGPT on the web, iOS, or Android, you’ll clearly notice more model options and voice capabilities that feel closer to real-time communication.
GPT‑5.1 and GPT‑5: Answer quality and availability advancing in parallel
According to the official statement, ChatGPT is gradually applying improvements to GPT‑5.1’s responses and rolling them out worldwide across web, iOS, and Android to various plans in stages. Meanwhile, GPT‑5 is also being opened up in batches on ChatGPT’s web, mobile, and desktop apps, covering plans such as Free, Plus, Pro, and Team; Enterprise and Edu will also gain access soon.
The model picker becomes “more controllable”: manually switching between GPT‑5 and Thinking mode
For paid users, ChatGPT’s model selector supports manually choosing GPT‑5 or GPT‑5 Thinking, making it easy to switch between “fast results” and “more reliable reasoning.” Pro and Team also get GPT‑5 Thinking Pro, which spends more time thinking but is better suited to complex tasks; additionally, due to GPU resource constraints, 4.5 is currently available only to Pro users.
“Show more models” toggle: enable hidden options all at once
ChatGPT’s web settings now include a “Show more models” toggle; once enabled, you can add models such as o3, o4-mini, 4.1, and GPT‑5 Thinking mini, making it easier to choose based on task intensity. At the same time, 4o has returned by default to the model selector for all paid users, making ChatGPT smoother to use for everyday writing, conversation, and multimodal tasks.
Voice upgrades: video, screen sharing, and image upload rolling out gradually
In the latest mobile apps, ChatGPT Advanced Voice is gradually adding support for real-time video, screen sharing, and image uploads, and is expected to reach all Team users and most Plus and Pro users in the short term (excluding some regions). Note that video and screen sharing have daily usage limits, and image uploads are also constrained by ChatGPT’s visual message limits—making them well-suited for “talk while showing” demos and remote assistance scenarios.
Deep Research and GitHub integration: closer to real-world workflows
ChatGPT Deep Research’s integration with GitHub is now available globally to Team users, and is beginning to roll out gradually to Plus and Pro (some regions not included for now). If you use ChatGPT for code reading, change summarization, or information consolidation, this kind of integration can reduce copy-and-paste overhead and connect “conversation” more naturally into your project collaboration workflow.
