Codex
Codex changelog
New features, models, price and usage-limit changes of Codex, newest first, each linked to its official source. Resets have their own page.
22 updates in the last 30 days · Latest: Codex Removes 5-Hour Limits for Pro Plans (Nov 8, 2026)
November 2026
Limit increase
Codex Removes 5-Hour Limits for Pro Plans
All Pro plans remove 5-hour limits. Compared to Plus, Pro 100 provides 10X usage, Pro 200 provides 20X, and Pro 500 provides 50X, along with access to ultrafast.
New model
Codex launches cyber-specialized model
Our latest cyber-specialized model is out. It’s available through Daybreak Red access.
Improvement
https://bsky.app/profile/thsottiaux-bot.eurosky.social/post/3mzbxpdjolw2d
Day 28/ Banked 2/2
October 2026
Limit increase
Codex doubles usage limits for Pro and Business Premium plans
All usage limits are doubled for Pro 100, Pro 200, Pro 500 tiers and Business Premium, effective today.
Improvement
Codex adds voice note and Siri texting features
Codex now lets you send voice notes and use Siri commands like “Hey Siri, text Iggy to …” for easier sharing. Available to all users immediately.
New model
Terra is back in Codex with 2X the usage of 6.1 Sol
Codex has brought back the Terra model with enhanced writing, personality, and computer use capabilities, offering 2X more usage than 6.1 Sol across API and subscriptions.
Improvement
Codex gets faster connects, a cache fix and a CLI update
Day 19 brings faster call connections for dots, mid-conversation reasoning changes that no longer invalidate the cache, and an updated Codex CLI.
Improvement
Codex Decisions API is generally available
Day 18 of Codex 28 days: the Decisions API is generally available, OpenResponses joined the Linux Foundation, performance improvements shipped and Uber is going all-in on Codex. The reset announced the same day landed on Oct 23.
Improvement
Codex Feature Included in Pro Plan
Codex has an improvement. A feature is fully included in the Pro plan for Pro plan users and does not draw on weekly usage.
Improvement
OpenAI Models and Codex Coming to Google Cloud
OpenAI announced that its frontier models and Codex agentic workflows are coming to Google Cloud's Model Garden for enterprise customers.
Improvement
Codex adds 12‑hour API cache and permanent user cache
Codex now caches API responses for 12 hours and keeps cache forever for logged‑in users across Codex & Work, with no penalties for revisiting old conversations.
Improvement
Codex adds message editing in desktop app
Codex now lets users edit previous messages within a thread in the desktop app, a long‑requested quality‑of‑life improvement.
Limit increase
Codex Increases Effective Usage via Better Caching
Codex improved effective usage by reusing context across subagents, retaining cached context longer, and optimizing request routing for higher cache hit rates.
Improvement
Codex Adds Nine New Presets
The nine new Codex presets were created with native-speaking actors in French, German, Hindi, Japanese, and Korean.
Improvement
Codex Launches Hardware Projects Community Showcase
OpenAI built a new site for users to explore and share community hardware projects created with Codex.
Improvement
Codex update cuts capacity errors
Codex improvement reduces the disruptive "Selected model is at capacity" errors.
Improvement
Codex Reduces Model Capacity Errors
Codex added compute and optimizations to significantly reduce 'Selected model is at capacity' errors and allow requests to auto-continue when capacity issues occur.
Improvement
Codex lists Composer predictions (beta) among its plan features
Codex added Composer predictions (beta) as an available feature on its pricing and plans page.
Discontinued
GPT-5.5 retires from Codex on Oct 14, 2026
GPT-5.5 will retire from Codex for users with ChatGPT sign-in on October 14, 2026, at noon PDT (19:00 UTC). OpenAI API and Codex sessions authenticated with an API key are not affected.
Improvement
Codex and ChatGPT Desktop App Storage Optimization
Optimizations to state storage on disk for the chatgpt and codex desktop app reduce the amount stored on disk by 60%.
Improvement
Codex Pairs Annotations with WebMCP
Annotations paired with WebMCP give your agents more context to act on in Codex.
Limit increase
Codex Increases Context Length for GPT-6.1 Sol
The default context length for GPT-6.1 Sol has been increased to 500k when used through Codex & Work, up from 272k or an ~ 80% increase.
Improvement
Codex desktop app gets Composer predictions
Composer predictions are now available in the desktop app. They are included in Pro plans without consuming usage.
Improvement
Codex launches composer predictions in beta
Codex has launched composer predictions in beta for Pro users. The feature suggests your next message based on your conversation history and communication style.
Improvement
Codex Sandbox Mode on Windows
OpenAI launched a new Codex sandbox mode using Microsoft Execution Containers (MXC) on compatible Windows 11 devices.
Improvement
Codex makes steering instant and releases GPT-6.1 Sol Ultrafast
Steering is now instant for real-time course correction, alongside the release of GPT-6.1 Sol ultrafast.
Price change
GPT-6.1 Sol Ultrafast API pricing: $12 / $60 per million tokens
OpenAI announced GPT-6.1 Sol Ultrafast mode API pricing at $12 per million input tokens and $60 per million output tokens.
Improvement
GPT-6.1 Sol Ultrafast comes to Codex, the API and ChatGPT Work
Ultrafast is launching for GPT-6.1 Sol across API, Codex, and ChatGPT Work, delivering near-Astra intelligence at up to 8x faster speeds.
Improvement
Codex Cloud Re-shipped
Codex quietly re-shipped Codex Cloud as part of Day 3 of its 28-day release cycle.
Improvement
Codex makes "Approve for me" free and adds a Decisions API
"Approve for me" (auto-review) is now included and does not use usage, alongside a simplified API, integrated meeting notes, and a new Decisions API.
Improvement
Codex Launches Decisions API
Codex has launched the Decisions API, introducing a realtime general classifier that opens up new possibilities for improving the user experience.
Improvement
OpenAI API Rate Limit Tiers Simplified
OpenAI simplified its API rate limit tiers from five to three (Build, Launch, Grow), making it easier to qualify for the highest Grow tier with $500 in total API payments.
Improvement
Codex Auto-review is now free for ChatGPT sign-ins
Auto-review is now free for all users signed in with a ChatGPT account. Turn it on in Settings > Permissions > Auto-review; it improves on the default sandbox setting, where you approve every action and decision fatigue sets in.
Improvement
Codex runs GPT-6 Astra and Sol about 50% faster by default
Default speed has been optimized to be ~50% faster across GPT-6 Astra and GPT-6.1 Sol for subscriptions across all products and partners using Sign in With ChatGPT.
New model
GPT-6.1 Sol Launched in ChatGPT Work and Codex
GPT-6.1 Sol is now available in ChatGPT Work and Codex (not Chat). For Enterprise and Edu, it is off by default and requires administrator enablement. Typical GPT-5.6 Sol task usage updated to 2-15 credits.
Discontinued
GPT-5.5 retires from ChatGPT and Codex on Oct 14, 2026
GPT-5.5 will be retired from ChatGPT, ChatGPT Work, and Codex across all plans on October 14, 2026. The OpenAI API will not be affected by this change.
September 2026
New model
OpenAI introduces GPT-6.1 Sol
GPT-6.1 Sol is available in Codex. Per OpenAI's Codex pricing page it uses 50 credits per 1M input tokens, 2.5 per 1M cached input and 250 per 1M output; in the API, cached input costs $0.10 per 1M tokens.
Limit decrease
Codex Pro 200 drops from 20x to 10x Plus usage; current subscribers keep 20x until Oct 29
Pro 200 reopened to new subscribers at 10 times ChatGPT Plus usage (was 20 times); the price stays $200/month. Eligible existing subscribers keep 20 times through October 29, 2026, then move to 10 times.
Price change
Codex Pro 500 plan: $500/month, 25x Plus usage with Ultrafast
OpenAI's new Pro 500 plan has the highest usage allowance, 25 times ChatGPT Plus, and is the only Pro plan with Ultrafast in Codex and ChatGPT Work. Available now.