Commentary | Codex Update

Codex plugin spotlight: @shuieng's Research Figure Kit keeps figures editable. It's an open-source (MIT) Codex plugin for paper and thesis figures. Instead of a flat AI image you have to re-prompt, it keeps the .drawio source, so fixing one arrow or label is a direct edit. 4 skills inside: - drawio: architecture, UML, flowcharts - drawio-academic-skills: paper-ready workflow and mechanism figures - drawio-poster: service-flow posters - scientific-visualization: Matplotlib, Seaborn or Plotly plots, profiling your data read-only first Install: git clone github.com/yss-off/research-… cd research-figure-kit codex plugin marketplace add . codex plugin add academic-figure-skills@research-figure-kit The core works offline. The live Draw.io backend is optional.

Known bug: Codex on iPhone says "No tasks yet." Your chats aren't gone. About 10 users across 3 GitHub issues hit it on ChatGPT iOS 1.2026.300 and .307, with Mac and Windows/WSL hosts. Remote still connects and new tasks still run, but the task list comes up empty. Workarounds for now: - Tap search: existing chats still open from there - Sign out and back in on the phone: the list comes back (until the next force-quit) - Push notifications and the Live Activity still open tasks Fair ask to OpenAI: fix the list on app relaunch and flag it in the iOS release notes. github.com/openai/codex/issu…

One flag swaps your whole Codex setup. Copy this: # ~/.codex/deep-review.config.toml model_reasoning_effort = "high" approval_policy = "on-request" Then run: codex --profile deep-review codex exec --profile deep-review "review this change" How it works: - Codex loads your normal ~/.codex/config.toml, then layers the profile file on top - So the profile only needs the keys that differ - One file per setup (deep-review, quick-fix, ci...) Gotcha: since CLI 0.134.0, --profile ignores old [profiles.name] tables in config.toml, and profile = "name" no longer works. Move them into ~/.codex/name.config.toml and delete the old lines. Docs: developers.openai.com/codex/…

Codex CLI v0.167.2 is out. Small patch, one change, for Google Cloud users. What's in it: - Codex now sends tool declarations incrementally by default on Google Cloud Vertex AI, instead of resending the full tool list every request - Same idea as v0.167.1's Bedrock flag, but this time it's on by default - Per the release notes, opt out with `incremental_tools_legacy_gcp = false` under [features] in config.toml Not on Vertex AI? Nothing changes for you. Update: npm i -g @openai/codex@latest github.com/openai/codex/rele…

Codex tip: audit a repo's .codex/ config before you trust it. agentaudit, open-sourced by Manifold Security, reads a repo's Codex, Claude Code and Cursor config and lists what it can do: - commands it runs (hooks, MCP servers, credential helpers) - env values it passes to servers - approval and sandbox changes - network redirects like proxies or API base URLs It only reads config. It never runs it. go install github.com/manifold-security… agentaudit github.com/OWNER/REPO GitHub URLs scan without cloning. Needs Go 1.26+, Apache-2.0. github.com/manifold-security…

Codex desktop 26.1103: 4 known bugs, 4 workarounds. 1. macOS: SSH remote projects don't reconnect at launch (#60047, 5 reports, still on build 25894) Workaround: Settings > Connections, re-enable the SSH connection 2. Ubuntu: Send stays grayed out at 0% weekly allowance even with credits left (#60445) Workaround: open Settings > Usage, or fully quit and relaunch 3. Windows: WSL reconnect loop keeps stealing focus from other apps (#60204) Workaround: uncheck "WSL: Ubuntu" under WSL connections 4. Linux: Enter and Send do nothing on build 26.1103.32239 (#60059) Workaround: roll back to 26.1027.51404 Fair ask to OpenAI: per-build desktop release notes (#60384), so users know which update fixes what. github.com/openai/codex/issu…

6 Codex CLI changes just merged, not released yet: 1. Cloud chats: pick the model and reasoning effort from the TUI (#60651) 2. Archive and restore cloud threads (#60635) 3. SSH or headless Linux: device-code sign-in for gateway providers, no local browser needed (#60662) 4. MCP tools marked readOnlyHint no longer ask for approval in auto mode (#60638) 5. Narrow terminals: composer footer shortens hints instead of clipping them (#60680 to #60683) 6. Safer repo checks: Codex skips submodule worktree scans that could run git clean filters (#60669) No release date yet. Watch for them here: github.com/openai/codex/rele…

Codex CLI v0.167.1 is out. Small patch, one change. What's in it: - New opt-in flag `incremental_tools_legacy` (off by default, marked under development) - It only applies to Amazon Bedrock providers (amazon-bedrock, amazon-bedrock-runtime) - With it on, Codex sends tool declarations as incremental updates instead of resending the full tool catalog on every request If you're not running Codex on Bedrock, nothing changes for you. No need to rush this one. Update: npm i -g @openai/codex@latest github.com/openai/codex/rele…

Known bug: Codex dictation can leave other apps stuck quiet. Your mixer says 100%, but browser audio stays low after dictating in the desktop app. What's been reported: - Linux: the dictation helper ducks other streams to 25%. If a stream ends before dictation does, it never gets restored, and it compounds (1.0, then 0.25, then 0.0625). Confirmed on Debian 13 and Arch. - Windows: Edge audio left at -18 dB with every slider at 100%, traced to dictation-ducking.exe. Linux workaround (PipeWire), reset the app's stream: wpctl status wpctl set-volume <stream-id> 1.0 Fair ask, OpenAI: always restore the original volume when dictation ends, even if the stream is gone. github.com/openai/codex/issu… github.com/openai/codex/issu…

Unconfirmed: GPT-6.2 Astra is not an announced Codex model. Posts today say `codex --model gpt-6.2-astra` is rolling out. What we can verify right now: - No announcement from OpenAI's official accounts - The official Codex models page lists Astra, GPT-6.1 Sol, GPT-6.1 Terra and GPT-6 Luna - The latest Codex CLI on npm is still 0.167.0 Before pinning a new model in config.toml or scripts, run /model in the CLI to see what your account can actually use. If this is a quiet test, a one-line note from OpenAI would save a lot of guessing. developers.openai.com/codex/…

Built with Codex: @warengonzaga made Codex a Suno music producer. - Writes original lyrics, builds Style prompts, preps faithful covers - Checks the full brief before spending Suno credits - Learns your style per project, no one-size preset Install: codex plugin marketplace add warengonzaga/suno-music-producer Heads-up: Suno's MCP access may be invite-only. Without it, the skill still preps lyrics and prompts. github.com/warengonzaga/suno…

Now live: one dot, multiple computers. You pick which machine runs each task. For Codex users, that means your dot can run a local Codex task on a second Mac or a build box while you keep coding on your laptop. How to set it up: 1. Open your dot's profile in the ChatGPT app on each computer 2. Under Computers, select Allow, then Allow access Each thread uses one computer at a time, and the ChatGPT app has to stay open on that machine. Connecting a computer to Codex doesn't give your dot access by itself. Docs: learn.chatgpt.com/docs/dots/…

Now live: the Codex edition ModRetro Chromatic on OpenAI Supply Co. What you get in Codex: the ModRetro Chromatic plugin, so you can design levels, animate sprites and playtest your game on the handheld. The shop also lists a ModRetro Codex cartridge and a Codex edition case. No price showed on the shop page when we checked, so confirm at checkout. Following up on yesterday's teaser. Shop: openai.com/supply/shop/

3 Codex CLI v0.167 shortcuts worth saving: 1. codex agents --no-daemon Opens the agent command center. Unread results are now marked. 2. Ctrl+L Clears the terminal display. Earlier messages stay available with Ctrl+T. 3. Chat won't resume after a provider switch It stays visible now. Open its transcript and prep an editable prompt to continue in a new chat. Update: npm i -g @openai/codex@0.167.0 Notes: github.com/openai/codex/rele…

Nov 5 Terra spike: ~30% of one user's week, gone. One Business user logged every GPT-6.1 Terra call and did the maths (GitHub issue 59690): - Nov 1 to 4: 0.32% of the 5-hour limit per call - Nov 5: 3.03% per call, about 9x more, with ~99% of input still shown as cached - 73 calls on Nov 5 moved the weekly meter 34 points. At the old rate that's about 2 to 3 Their read: cached input on Terra was briefly billed at full price. By Nov 6 the same CLI build (0.166.0) was billing normally again, so the fix looks server-side. Not a one-off. Issue 59639 links 16 reports of unexpected quota burn opened Nov 2 to 5, across CLI, desktop and IDE. Check yours: compare your weekly meter in /status with what you actually ran on Nov 5. Fair ask to OpenAI: confirm what happened on Nov 5 and restore usage for accounts that were overcharged. github.com/openai/codex/issu…

Known bug: Codex desktop refuses to schedule job monitoring. OpenAI's Scheduled tasks docs list "checking a long-running operation until it finishes" as a use for a schedule inside a chat. Since Nov 5, users on Mac and Windows report the desktop app declining exactly that, even for a GitHub PR or a remote training job that has nothing to do with Codex. What's going on, per GitHub issue 59546: - The app's built-in instructions say not to create an automation that tracks a job or task through a milestone, failure or completion - The assistant reads that as covering outside jobs too, and stops before it ever calls the scheduler - Its fallback is to keep checking in the current turn, which ends when the turn does Reported on desktop builds 26.1029 and 26.1103. One user says other chats still schedule fine, but no reliable workaround has been confirmed yet. Fair ask to OpenAI: scope that instruction to Codex's own tasks, or say in the docs that external job monitoring isn't supported. github.com/openai/codex/issu…

November 9: Codex x ModRetro Chromatic drops on Supply Co. @OpenAIDevs just posted a "Tomorrow." teaser. The video shows an OpenAI-branded handheld box, the OpenAI and ModRetro Chromatic logos, and a clip calling it the Codex collab. This is the same partnership ModRetro announced in September: Chromatic hardware plus Codex for making games. The teaser gives no price or specs, and the item is not on the Supply Co. shop yet. If you want one, watch Supply Co. on November 9: openai.com/supply/shop/

Known bug: desktop projects vanish after you switch accounts. After the 26.1103 Codex desktop update, switching ChatGPT accounts can empty the local Projects sidebar on Windows and Mac. See GitHub issue 59895. Your chats stay under Recents, and the project folders are still on disk. The work itself is not deleted. Workarounds people have reported: 1. Fully quit the app with File > Quit ChatGPT (closing the window is not enough, because it can stay in the tray), then reopen it while already signed in. This brought the list back for several reporters. It does not work for everyone, and the empty sidebar can return after the next account switch. 2. Add one of the missing projects again, using the exact same name and folder. One reporter said every missing project then reappeared, including the original with its chat history. Delete the empty duplicate afterward. Fair ask to OpenAI: keep local projects visible across accounts on the same machine, as they were before this update, or ship a supported restore path so people do not have to guess whether their work is gone. github.com/openai/codex/issu…

Codex this week: 7 changes, 5 user pain points, all here. Nov 1 to Nov 8, checked against OpenAI's posts, changelog, status page and GitHub. Releases 1. CLI v0.167.0 (Nov 6): agent command center via codex agents --no-daemon, switch backends without losing your draft, opt-in system proxy, and saved chats no longer vanish when you change model providers. 2. GPT-Image-2.6 came to Codex and the API (Nov 4). OpenAI says the new Flare and Sunburst API models generate images 50% faster. 3. GPT-6 Cyber launched Nov 5, in Codex and the API for approved Daybreak Red orgs only. 4. ChatGPT iOS 1.2026.300 (Nov 2): follow-ups steer the running task by default, and you can manage attachments in the full-screen composer. 5. Dots on iOS: reply inline from a notification, share straight to your dot, open it from the Action button. Tibo says Terra and Astra will power dots too. Limits and pricing 6. Banked reset 2 of 2 landed Nov 1, the last day of Tibo's 28-day run. Redeem it in Usage settings or with /usage before it expires. 7. Tibo spelled out current Pro rates: Pro 100/200/500 = 10x/20x/50x Plus usage, no 5-hour cap on any Pro plan, ultrafast on Pro 500 only. Not a new increase. He also asked whether a Pro 50 tier would make sense. Just a question for now. Incidents - Nov 5: errors in ChatGPT Work and dots for about 1h40m until a mitigation went in. What users are saying - "Selected model is at capacity" with plenty of quota left. The loudest complaint of the week, on X and in several GitHub issues (#58824, #58858, #59059), across CLI and desktop, Plus and Pro. Several reporters say switching models doesn't help. No official explanation yet. - Local projects vanish from the desktop sidebar after the 26.1103 update when you switch accounts (#59895, Windows and Mac). Chats and folders are still on disk. One user got them all back by re-adding a project with the exact same name and folder, then deleting the duplicate. - Attaching an .xlsx leaves Send spinning (#58660 Windows, #59362 Mac). Two reporters got around it by typing the file's local path in the message instead. - Typing lag in dot chats in the Mac desktop app (#59301). Several users say the same dot runs fine in Chrome. - Pushback on December's shared ChatGPT + Codex pool continues on GitHub (#58742, #58655). What people are asking OpenAI for: a status note or a clearer error when accounts with quota left get "at capacity", local projects shared across accounts on one machine again, and per-plan token numbers before December. Community shoutout @turushan built Codex Whip, a free, open-source Mac menu bar app. When Codex CLI stops on a capacity error or usage limit, it waits, then continues the same session with the same model: codexwhip.com Tip Reopen your latest CLI session in this folder, no picker: codex resume --last Sources: openai/codex releases and issues on GitHub, the ChatGPT & Codex changelog, status.openai.com, @OpenAI, @OpenAIDevs, @thsottiaux and @_samirism on X

Pro 100 = 10x Plus usage, no 5-hour cap. Tibo from OpenAI's Codex team just spelled out the current Pro multipliers vs Plus: - Pro 100 = 10x - Pro 200 = 20x - Pro 500 = 50x, plus access to ultrafast This is a clarification of today's rates, not a new increase. Ultrafast stays on Pro 500 only. Buying extra credits on Pro 100 or Pro 200 does not unlock it. Check your tier in Settings > Usage, or run /usage in the CLI. Help center: help.openai.com/en/articles/…