OpenAI adds push notifications, /fork command to Codex iOS app
OpenAI has updated Codex in ChatGPT for iOS with push notifications for turn completion, a new /fork command, and fixed a bug preventing Sign in with Apple users from linking the feature. The company continues its pattern of Thursday feature releases for the desktop app.
OpenAI adds push notifications, /fork command to Codex iOS app
OpenAI has updated Codex in ChatGPT for iPhone and iPad with push notifications for turn completion and new commands, according to the company's updates this week.
Bug fixes and new features
OpenAI fixed a bug that prevented users with Sign in with Apple accounts from setting up Codex on iOS. The issue incorrectly displayed a Google sign-in form instead of the appropriate authentication method.
The update adds push notifications that alert users when a task on Codex for Mac completes or requires input. This eliminates the need to actively monitor task progress on the desktop app.
OpenAI also added the /fork command to the iOS app, with /side planned for future release. The update includes improved reconnection UI, more compact conversation views matching the desktop app, and enhanced diff viewing with an option to open full files.
Regular update cadence
OpenAI has established a pattern of releasing major Codex desktop features on Thursdays. The company is expected to announce new desktop capabilities today, continuing several weeks of Thursday releases.
Codex launched on Mac in February 2026 after initially being available only as a command line interface tool. The iOS integration arrived last week, bringing the agentic coding tool to iPhone and iPad through ChatGPT.
Recent Codex developments
Last month, Codex gained the ability to use Mac apps without controlling the cursor, allowing users to continue working while Codex executes tasks. OpenAI recently introduced a Codex-specific subscription tier and released GPT-5.5, which upgraded capabilities across ChatGPT and Codex. The company also launched Images 2 for image generation.
What this means
The rapid iteration on Codex shows OpenAI treating it as a distinct product line rather than just a ChatGPT feature. Push notifications address a key usability gap for mobile users who initiate tasks remotely. The consistent Thursday release schedule suggests OpenAI is running focused sprint cycles on Codex development, likely responding to competition from other AI coding tools. The /fork command addition indicates OpenAI is building more advanced version control capabilities directly into the tool.
Related Articles
Simon Willison's Pelican Benchmark Shows GPT-6 Astra Outperforming GPT-5.6 Sol at Every Reasoning Level
Developer Simon Willison ran his signature 'pelican riding a bicycle' SVG test on newly-accessed GPT-6 Astra across five reasoning levels, comparing results against GPT-5.6 Sol, Terra, and Luna. Even Astra's lowest reasoning setting reportedly beat every Sol output, though Astra costs roughly twice as much per token.
Gemini Overlay on Android Adds Minimize Button for Multitasking Bubble
Google is widely rolling out a new Minimize button for the Gemini overlay on Android, which collapses conversations into a floating bubble users can drag or tap to expand. The feature currently supports six fixed positions and has a reported bug that resets bubble placement after each minimization.
OpenAI Lists GPT-6 Astra Pro on OpenRouter: Same Model, Higher-Compute Reasoning Mode
GPT-6 Astra Pro, now listed on OpenRouter, is the existing GPT-6 Astra model configured to run with reasoning.mode set to 'pro' for higher-quality output on complex tasks. It carries a 1M-token context window and tiered pricing from $5/$25 to $20/$100 per million input/output tokens depending on the serving tier.
OpenAI's GPT-6 Astra Cuts Hallucinations, But Indirect Prompt Injection Attacks Still Succeed 8.5% of the Time
OpenAI's new GPT-6 Astra model shows major improvements in hallucination rates and jailbreak resistance over predecessor GPT-5.6 Sol, according to OpenAI's system card. However, indirect prompt injection attacks hidden in documents still succeed 8.5% of the time in external testing by Gray Swan, down from 27% but still above rival Claude Opus 5's 4.8% rate.
Comments
Loading...