AI Tool Updates

Claude Code Newly added 2.1.267 New maxEffortLevel Setting Caps Reasoning Effort Across Providers

Setting maxEffortLevel, either top-level or per model under modelSettings, caps the reasoning effort level across every provider including Bedrock, Vertex, and Foundry; users can still choose a lower level.

What changes If you need to manage cost or response speed across a team, you can set one organization-wide cap instead of asking everyone to configure it session by session.Source →

Claude Code Newly added 2.1.267 Claude Now Suggests /copy When Clipboard Commands Fail in the Sandbox

When clipboard commands like pbcopy fail inside the sandbox, Claude now suggests using /copy instead.

What changes When copying suddenly stops working in the sandbox, you no longer have to guess why; Claude points you straight to /copy instead.Source →

Claude Code Newly added 2.1.267 Faster --resume for Sessions With Many Bash Tool Calls

Resuming a session with many accumulated Bash tool calls now renders its first screen faster.

What changes The longer a session's Bash history, the less you now wait after --resume before you can pick up where you left off.Source →

Claude Code Newly added 2.1.267 Fixed Keystrokes Occasionally Lagging Behind the Spinner

Prompt input responsiveness was improved: keystrokes no longer occasionally wait a frame behind spinner or streaming repaints.

What changes Typing while Claude is streaming a long answer no longer occasionally looks laggy or glitchy on screen.Source →

Codex Newly added rust-v0.154.0 GPT-6 Astra Is Now a Model Choice in Codex

GPT-6 Astra is now available in Codex's model picker and in the Amazon Bedrock catalog.

What changes For harder tasks, you can just switch to Astra from the model picker without any extra setup.Source →

Codex Newly added rust-v0.154.0 New --worktree Flag Gives Each Session Its Own Checkout

The --worktree flag or /worktree command creates an isolated checkout for a new or forked session, which you can later browse and resume.

What changes When you're running several tasks at once, you no longer have to juggle branches in one folder; each session gets its own fully separate directory.Source →

Codex Newly added rust-v0.154.0 Answer Codex's Mid-Task Questions Without Losing Your Draft

When Codex asks a question mid-task, you can answer inline with a suggested choice or custom text, without losing whatever you were drafting in the main input.

What changes When Codex asks which way to go, you can just answer on the spot and keep going, no need to clear or save what you were typing.Source →

Codex Newly added rust-v0.154.0 MCP Connections Now Prompt Re-Login When Token Refresh Fails

MCP connections now coordinate OAuth token refreshes and surface a login prompt when a refresh fails, instead of silently retrying a rejected tool call.

What changes Instead of a connected MCP tool silently failing for no visible reason, you now get a clear prompt to log back in and keep going.Source →

ChatGPT Newly added ChatGPT Voice Can Now Answer With GPT-5.6 or GPT-6 Astra

ChatGPT Voice can now use GPT-5.6 or GPT-6 Astra when it needs to search or reason through harder questions, with the same model and reasoning-effort controls as text chat. Plan-based usage limits were also simplified.

What changes You can now expect voice answers to complex questions to be as capable as text chat. Usage time still depends on your plan, so it's worth checking your limit in settings.Source →

ChatGPT Newly added Share Files and Folders From Your ChatGPT Library

You can now share files and folders from your ChatGPT Library, choose who can access them, and work with shared content directly inside conversations.

What changes Instead of re-attaching files for a colleague every time, you can share a Library folder once and keep referencing it across future conversations.Source →

Codex rust-v0.153.4 Codex adds GPT-6 Astra to its model picker and Amazon Bedrock

Codex added OpenAI's newer GPT-6 Astra model to its bundled model picker and to Amazon Bedrock's catalog, along with refined in-app guidance for Astra-specific features like asynchronous questions.

What changes You can pick Astra straight from Codex's model picker without extra setup, and teams deploying through Bedrock can now use the same model.Source →

Notion Notion lets you ask your agent to suggest document edits

In Notion, you can now ask an AI agent to propose edits to a document instead of changing it directly, so a person reviews and decides whether to accept each suggestion.

What changes If you've hesitated to hand a document to an agent for fear it would rewrite things outright, you can now just review its suggested edits and approve them yourself.Source →

Notion Notion simplifies its AI model picker

Notion simplified the list of AI models you can choose from when using its AI features.

What changes If you were never sure which model to pick, the shorter list means one less decision before you can just get started.Source →

Notion Notion's AI Meeting Notes can now kick off Custom Agents

When Notion's AI Meeting Notes finishes, it can now automatically kick off a Custom Agent to handle follow-up work.

What changes The task of reading meeting notes and manually entering action items after every meeting can now be handed off to an agent you've already set up.Source →

Notion Notion agents get calendar tools

Notion agents now have tools to check and adjust calendar events directly.

What changes Instead of checking back and forth with a calendar to coordinate a schedule, you can let an agent handle it.Source →

Claude Code 2.1.265 Claude Code caps saved tool results at 1 GB

Claude Code now caps tool results saved to disk at 1 GB, and the in-conversation preview notes when a saved file was truncated.

What changes Running a task that produces huge output no longer risks suddenly filling up disk space, and you're less likely to miss that a result got truncated.Source →

Claude Code 2.1.265 Claude Code fixes Windows sandbox blocking all file access

Fixed a Windows bug where running inside an AppContainer or restricted-token sandbox made Read, Write, and Edit refuse every file.

What changes If you were running Claude Code in a Windows sandbox environment, the situation where it couldn't read or write any file at all is now fixed.Source →

Claude Code 2.1.265 Claude Code shows a list of matches for slash commands typed mid-prompt

Typing a slash command in the middle of a prompt now shows matching commands as a list instead of a single suggestion; Tab opens it even outside fullscreen, and a plugin skill can be found by its bare name.

What changes When you're mid-prompt and can't quite remember a command's name, you can just pick it from the list instead of retyping.Source →

ChatGPT ChatGPT images get templates and sketch-to-image

ChatGPT's image generation moved to version 2.5 with sharper output and more precise editing. Users can now start from a template, turn a hand-drawn sketch into an image on mobile, or comment directly on a generated image to request changes.

What changes When it's hard to describe an idea in words, starting from a rough sketch or a ready-made template gets you to the result you want faster.Source →

Notion Notion adds a high contrast mode

Notion added a high contrast display mode to its settings.

What changes If low contrast made Notion hard to read, you can now turn this on in settings for a clearer display without a browser extension.Source →

Claude Code 2.1.261 Claude Code now shows why your org policy failed to load, right in /status

Added an "Organization policy" line to /status and claude doctor that explains why your organization's policy failed to load, such as a proxy not passing the endpoint through.

What changes Instead of guessing why company policy isn't applying, running /status now tells you the actual cause, like a proxy misconfiguration, so you can hand IT the exact problem.Source →

Claude Code 2.1.261 You can now raise the character limit on long command output

Added bashOutputMaxChars and taskOutputMaxChars settings to raise how much command and background-task output Claude receives inline before it's saved to a file, up to 128K characters.

What changes If long test logs or build output used to get cut off mid-way, raising this setting lets Claude see the whole thing at once instead of missing context.Source →

Claude Code 2.1.261 Long subagent prompts can now be passed as a file

Added --append-subagent-system-prompt-file to read the subagent system prompt from a file, for prompts too large to pass on the command line.

What changes If you build custom subagents, you can now hand over a long instruction set as a single file instead of trimming it to dodge shell-escaping issues.Source →

Claude Code 2.1.261 /skill-doctor shows which unused skills are eating your context

Added /skill-doctor to show which loaded skills go unused and what they cost in context, so you can prune them.

What changes If a session feels sluggish with several skills loaded, /skill-doctor can point out the unused ones so turning them off frees up context.Source →

Claude Code 2.1.261 /model picker now shows model names instead of raw IDs

Improved the /model picker and the VS Code model pill to show a model's name instead of its raw Bedrock, Vertex AI, or LLM gateway ID when Claude Code recognizes it.

What changes If you use Claude through Bedrock or Vertex AI, you can tell which model is selected at a glance instead of decoding a long ID string.Source →

Claude Code 2.1.261 Faster startup on Google Vertex AI when credentials are pre-set

Improved startup on Google Vertex AI when GOOGLE_APPLICATION_CREDENTIALS is set: API client creation no longer re-runs Google Cloud project discovery or spawns extra gcloud processes.

What changes If Claude Code felt slow to start on Vertex AI, this fix skips the repeated gcloud calls on every run, cutting the wait.Source →

Claude Code 2.1.261 Long streaming responses no longer slow down the display

Improved streaming performance: already-rendered blocks are no longer re-checked by layout on each update.

What changes Long responses that used to make the terminal stutter should feel smoother, cutting the perceived wait during long outputs.Source →

Claude Code 2.1.261 The dangerous rm -rf warning now catches more disguised forms

Improved the dangerous-rm safety prompt to also catch rm -rf on positional parameters and inside double-quoted sh -c scripts.

What changes rm -rf commands hidden in variables or nested scripts now trigger a warning before running, adding one more check against accidental deletion.Source →

Codex rust-v0.153.3 Codex now offers GPT-6 Astra through Amazon Bedrock

GPT-6 Astra was added to Codex's Amazon Bedrock model picker and catalog, for Mantle and Runtime global/US routes.

What changes Teams required to route models only through Amazon Bedrock can now pick GPT-6 Astra directly in Codex, with no separate approval or API key changes needed.Source →

Claude Code 2.1.259 Claude Code narrows what allowedMcpServers governs (check on upgrade)

The allowedMcpServers setting now governs only MCP servers that users add themselves. A specific managed-mcp.json server that this allowlist used to filter out will now load automatically on upgrade; use deniedMcpServers to keep it off.

What changes If you were relying on allowedMcpServers to block a specific managed server, it may come back on after upgrading. Check your deniedMcpServers setting.Source →

Claude Code 2.1.257 Claude Code now prompts before its first read outside your working directory

In auto mode, Claude Code now shows a one-time prompt before its first file read outside your working directories, with a new permissions.blockReadsOutsideWorkingDirectories option to block such reads entirely.

What changes If you run auto mode across multiple projects, this setting lets you prevent it from reading files outside the current one by accident.Source →

Claude Code 2.1.259 Claude Code preserves nested background subagent results

Results from a background subagent launched inside another background subagent are now saved to the parent subagent's transcript, so resumed sessions keep them and shared transcripts show that delivery happened.

What changes If you nest background agents inside each other, you're less likely to lose track of a sub-task's result when you resume the session.Source →

Claude Code 2.1.259 Claude Code's VS Code extension adds status filters to the session list

The VS Code extension's session list sidebar now has an Active quick filter and a status filter menu (Needs input, Working, Completed).

What changes When you're running several sessions at once, filtering down to the ones waiting on input cuts the time spent hunting for what's happening where.Source →

ChatGPT ChatGPT adds Zendesk and OneNote plugins

Zendesk and OneNote plugins were added to the plugin directory in ChatGPT and Codex. Users connect their own accounts to pull in support tickets or notes.

What changes When preparing a support reply, you no longer need to switch between Zendesk and ChatGPT — you can pull up the ticket history and draft the reply inside ChatGPT itself.Source →

ChatGPT ChatGPT Sites can now be shared with people outside your workspace

Eligible Site owners can now share a live ChatGPT Site with named people outside their workspace by email, without making it public. Invited viewers can view but not edit or republish.

What changes You can open up a company ChatGPT Site to one outside partner or client without making it public, just by entering their email in the Share panel.Source →

Codex rust-v0.153.1 Codex lets you use GPT-6 Astra via API without switching your default model

Codex now lets you configure GPT-6 Astra through the API without changing your default model or having it show up in the model picker.

What changes You can try Astra's performance on a specific task without touching your team's default configuration.Source →

Codex rust-v0.153.0 Codex TUI now reconnects without losing your session

Codex TUI sessions now automatically reconnect after an external app-server connection drops, keeping drafts and transcripts intact, with uncertain or queued submissions paused for review.

What changes A brief network drop is less likely to wipe out a session you were in the middle of.Source →

Codex rust-v0.153.0 Codex Guardian review history now survives restarts and forks

Codex's Guardian review history now survives conversation compaction, restarts, and user-created forks, while still respecting rollback boundaries and keeping subagent history separate.

What changes Even after compacting or restarting a long session, you keep a record of what was approved, so you can trace back what happened later.Source →

ChatGPT ChatGPT begins rolling out GPT-6 Astra

ChatGPT is rolling out GPT-6 Astra, with improvements in coding, research, computer use, and complex multi-step work, starting with a limited set of organizations. It's not yet generally available, with broader rollout planned over the coming days.

What changes Astra can pause a conversation on its own and ask you to review it if it suspects it may have misread your instructions. If ChatGPT suddenly stops mid-task, that may be this safeguard rather than an error.Source →

Codex Codex Vim mode now has undo and redo

In Codex's Vim mode you can undo with u and redo with Ctrl+R. Pasted content and attachments are kept as part of the draft.

What changes If you wiped a long prompt by accident, you used to start over. Now one undo brings back the whole draft, including pasted code and files.Source →

Codex Install Codex plugins from remote marketplaces via the CLI

The plugin CLI can list, install, and remove plugins from remote marketplaces.

What changes Instead of downloading and placing plugins by hand, you can install and clean them up straight from a marketplace with one command.Source →

Codex You can turn off Codex's automatic recaps

Set tui.auto_recap = false to turn off automatic recaps. The manual /recap command still works when you want one.

What changes If the recaps that appear during long sessions got in your way, you can switch them off and call /recap only when you want it.Source →

Codex Codex history now shows full patches and commands

TUI history now shows complete patches, input sent to background terminals, and each completed command on its own.

What changes You can retrace exactly what Codex changed and which commands it ran, in full rather than a clipped summary. That helps when you check back or trace a problem.Source →

Codex Codex warns about your usage limit earlier

Plus and Team users get an earlier warning once less than half of their allowance is left in a roughly five-hour usage window.

What changes Instead of hitting the limit mid-task, you learn when about half is left and can plan when to run heavier work.Source →

Claude Code 2.1.259 Claude Code Adds an Org-Wide Managed Setting for MCP Servers

Claude Code added a managedMcpServers setting, letting organizations roll out HTTP/SSE MCP servers to every user at once, using the same entry format as .mcp.json. Entries that specify a command to run are skipped for safety.

What changes Instead of configuring MCP servers for each team member individually, an admin can now set them up once for everyone.Source →

Claude Code 2.1.259 Claude Code Adds --permission-prompts none for Unattended Runs

Added --permission-prompts none for unattended headless hosts: anything that would normally prompt for confirmation is now automatically denied, while the active permission mode, including auto mode, keeps deciding everything else.

What changes When running Claude Code in a CI pipeline or on a server where no one can answer a prompt, this keeps the run from hanging on a confirmation it can never receive.Source →

Claude Code 2.1.259 Claude Code Now Recognizes GitLab MR Commands

Claude Code now recognizes glab mr create/merge/close/reopen/note/update commands, showing them as "MR !N" in the collapsed tool summary and refreshing the footer MR badge.

What changes If you work in GitLab instead of GitHub, merge request commands now show up in the interface just as clearly as GitHub PR commands already did.Source →

Claude Code 2.1.259 Claude Code's Plugin Validator Gets a --json Output Option

Added --json to claude plugin validate, producing a machine-readable validation report.

What changes If you want to wire plugin validation into a CI pipeline to automatically pass or fail a build, you no longer need to parse human-readable output to do it.Source →

Claude Code 2.1.259 Claude Code Speeds Up Terminal Resize and First Render for Long Responses

Claude Code now reuses text measurements, improving terminal resize and first-render performance for long responses.

What changes If you keep several long responses open and resize your window often, the stutter that used to cause is now reduced.Source →

Claude Code 2.1.257 Claude Code's default model switches to Fable 5.1

Claude Code's default model is now Fable 5.1 (claude-fable-5-1), with a 1M-token context window, priced at $10/$50 per million input/output tokens and $0.25 per million tokens for cache reads.

What changes You'll be working on the cheaper, faster model starting your next session with no setup needed, and the savings are biggest on long sessions that lean on cache reads.Source →

Claude Code 2.1.257 Claude Code adds a time format setting

A new timeFormat/timeZone setting lets you choose 12-hour, 24-hour, 24-hour UTC, or a custom format for turn-end and transcript timestamps.

What changes Useful if you work across time zones or share logs with a team and want timestamps in a familiar format.Source →

Claude Code 2.1.257 Claude Code adds a containment-escape guard to auto mode

In auto mode, attempts to fetch cloud metadata credentials, evade egress monitoring, or reach across tenants are no longer auto-approved, unless your environment explicitly marks them as expected.

What changes If you let auto mode run unattended, risky actions like leaking cloud credentials are less likely to slip through without a human checking first.Source →

Claude Code 2.1.257 New env var forces one model across all subagents

Setting CLAUDE_CODE_SUBAGENT_MODEL_FORCE applies CLAUDE_CODE_SUBAGENT_MODEL (or the main model) to every subagent, overriding any per-spawn or agent-definition model choice.

What changes If running many subagents with mixed models made cost or behavior hard to predict, this one variable pins them all to a single model.Source →

Claude Code 2.1.257 /effort can now change effort for just the current session

Adding s to /effort now changes the effort level for only the current session, matching how /model already works, and it resets on the next session.

What changes Handy for trying a different effort level in just one session without changing your default going forward.Source →

ChatGPT ChatGPT for Clinicians adds US public healthcare data search

Eligible US users of ChatGPT for Clinicians can now search nine public healthcare data sources, including clinical trials, medication information, Medicare data, and provider records, directly inside ChatGPT.

What changes When you need to look up a clinical trial or drug information mid-consultation, you can check it inside ChatGPT instead of switching to another site.Source →

Claude app Claude app now runs on Fable 5.1 and Mythos 5.1

The Claude app's default models are now Fable 5.1 and Mythos 5.1.

What changes You don't need to change any settings, your next conversation already runs on the cheaper, faster model.Source →

Claude Code 2.1.251 Claude Code Cuts CPU Usage During Turns

Claude Code improved CPU usage during turns in interactive sessions by cutting redundant UI re-renders.

What changes Running several long sessions at once now spins up your laptop fan less and drains less battery.Source →

Claude Code 2.1.251 Claude Code Cloud Sessions Now Name the Cause of Connection Drops

When a cloud session's network proxy drops a connection during a Bash command, the tool result now names the host and reason instead of only showing 'connection reset.'

What changes When you're running several long sessions at once, you'll spend less time hunting for which window dropped and why.Source →

Claude Code 2.1.251 /schedule Now Explains Why It Can't Attach Your MCP Servers

MCP servers configured in Claude Code can't be attached to cloud routines. Instead of a bare 'No MCP connectors' message, /schedule now explains why.

What changes You'll spend less time confused about why a scheduled routine won't pick up your MCP connector, since the reason is now spelled out.Source →

Claude Code 2.1.251 Claude Code Now Tells Claude Its Subagent's Messages Aren't From an Outside Session

Claude Code now tells Claude that a message from its own subagent comes from a worker inside this session, not from an unrelated Claude session.

What changes When you run multiple subagents at once, Claude now responds with a more accurate sense of where each message actually came from.Source →

ChatGPT ChatGPT's Browser Extension Now Works in More Browsers

The ChatGPT browser extension now supports Microsoft Edge, Brave, Opera, and Vivaldi. You can mention open tabs as context or ask the desktop app to complete a browser task.

What changes If you've been using a non-Chrome browser, updating the desktop app is all it takes to get the same features. Side chat only works on Edge, Brave, and Vivaldi, though; Opera supports tab mentions and browser control but not side chat.Source →

ChatGPT Create and share your own sticker pack in ChatGPT

Give ChatGPT a prompt or a photo and it generates a sticker set you can download or add directly to iMessage or WhatsApp.

What changes Handy for adding a custom sticker to something like a birthday message or a team announcement.Source →

ChatGPT See your ChatGPT voice conversation on the Lock Screen

On iPhone, Live voice conversation content now appears on the Lock Screen and in the Dynamic Island, so you can follow along without opening the app.

What changes You won't lose track of a voice conversation even while using another app or with your phone locked.Source →

ChatGPT Use website tools directly in ChatGPT's desktop browser

In the desktop app's built-in browser, ChatGPT Work and Codex can now use tools a website provides directly, without a separate connection. Selecting the arrow in the address bar shows what tools a page offers.

What changes You can use a website's own tools mid-task without setting up a separate integration first.Source →

ChatGPT Improved pronunciation help in ChatGPT

Ask how to pronounce a word or phrase, and you can now tap to hear it aloud and see a phonetic breakdown.

What changes Useful when you want to hear a foreign word or a tricky technical term spoken aloud.Source →

Claude Code 2.1.251 Claude Code adds hooks before and after a model switch

Claude Code adds PreModelSwitch and PostModelSwitch hook events, letting you block, confirm, or annotate a model switch. SessionStart resume hooks now also receive the session's staleness and the estimated cost of re-warming the prompt cache.

What changes Teams mixing multiple models can now automatically block or log switches to a given model, for tighter cost and quality control.Source →

Claude Code 2.1.251 Remote Control now streams a foreground subagent's work live

When watching a session via Remote Control, a foreground subagent's tool calls and results now stream live; background subagents still show status only, as before.

What changes When juggling several long-running sessions, you can now check exactly what a foreground subagent is doing from your phone or another device, cutting down the time spent figuring out which window was doing what.Source →

Claude Code 2.1.251 Claude Code adds a spend-limit bar to /usage

For developers behind a Claude apps gateway with spend limits, /usage now shows a spend-limit bar, and a new rate_limits.spend_limit field is available for status line scripts.

What changes If you work under a company-imposed spend limit, you can now see how close you are to it with a single command.Source →

Claude Code 2.1.251 Claude Code's /cost now shows prompt-cache hit rates

Claude Code's /cost command adds a per-session prompt-cache line (hit ratio, misses, tokens re-cached, warm/cold), plus a matching prompt_cache object for status line scripts.

What changes You can now see, in numbers, exactly how much keeping a session warm is actually saving on cost.Source →

Claude Code 2.1.251 claude --help now lists attach, logs, stop, and more session commands

claude --help now lists the attach, logs, stop, respawn, and rm commands, and the resume message for a running background session now names the exact claude attach <id> command to run.

What changes When you need to reattach to or check logs from a background session, you can just copy the exact command Claude Code shows you instead of hunting for the right syntax.Source →

ChatGPT ChatGPT now links multiple Google accounts at once

ChatGPT can now connect multiple Gmail, Google Calendar, and Google Contacts accounts at once, bringing personal and work accounts into the same conversation, across web, desktop, iOS, and Android on Plus, Pro, Business, and Enterprise plans.

What changes When you need to check schedules or find emails across a personal and a work account, you can now ask ChatGPT about both in the same conversation instead of logging out and back in each time.Source →

Claude Code 2.1.248 Claude Code adds a --restricted flag

Turning on --restricted (or setting CLAUDE_CODE_RESTRICTED=1) removes the built-in tools that run commands or code plus WebFetch (unless explicitly named in --tools), confines file tools to the working directory, refuses bypassPermissions, and ignores user, project, and local settings files.

What changes When working with untrusted code or an external repo, this one flag locks down command execution and file access without having to hand-tune permission settings each time.Source →

Claude Code 2.1.248 Set prompt cache TTL per agent

Agent frontmatter now supports experimental.cacheTtl ("5m" or "1h"), letting you set a per-agent prompt cache TTL used when no subagent-level TTL is configured.

What changes If you mix long-running background agents with ones you call briefly and often in the same session, you can now tune each one's cache lifetime separately to cut down on avoidable cache misses.Source →

Claude Code 2.1.248 Self-hosted runners can register with a custom label

claude self-hosted-runner --client-label <name> (or the SELF_HOSTED_RUNNER_CLIENT_LABEL variable) now lets you override the label a runner registers with, which previously defaulted to the hostname.

What changes If you run several servers with similar hostnames and couldn't tell them apart in the runner list, you can now give each one a name that's actually easy to recognize.Source →

Claude Code 2.1.248 Claude Code now tells you why your settings didn't load

When server-managed settings fail to load, Claude Code now shows a startup warning, and /doctor and /status explain why the load failed or why settings weren't fetched at all (Bedrock, Vertex, third-party providers, a custom ANTHROPIC_BASE_URL, etc).

What changes If org-managed settings seem like they're not taking effect, you can run /doctor once and see the actual reason instead of guessing.Source →

ChatGPT ChatGPT's temporary chat can now use memory and custom instructions

ChatGPT's temporary chat mode, which doesn't save to history, can now optionally use your regular memory, plugins, and custom instructions, and you can choose to save a temporary chat afterward.

What changes Even when you don't want a conversation saved to your history, ChatGPT can still use context it already knows about you, like your tone preferences.Source →

Claude Code 2.1.247 Claude Code drafts a feedback report the moment something goes wrong

Claude Code added a SendFeedback tool: when something goes wrong in a session, Claude drafts a feedback report you can review and send via /feedback, toggleable with the feedbackDrafts setting.

What changes Instead of writing up what happened from scratch every time you hit a bug, Claude drafts the report for you based on what just happened in the session.Source →

Claude Code 2.1.247 Claude Code adds an auto-mode tip on Bash permission prompts

Claude Code now shows a tip on Bash permission prompts pointing to auto mode, with a one-keystroke "Yes, and switch to auto mode" option.

What changes If you're tired of approving the same kind of command over and over, this tip lets you switch straight into auto mode in one keystroke.Source →

Claude Code 2.1.247 Claude Code adds a cost-optimize command to /claude-api

Claude Code added /claude-api cost-optimize, which profiles an existing project's Claude API spend and walks through cost levers, caching, token hygiene, batching, effort level, model choice, one measured change at a time.

What changes When your API bill runs higher than expected, instead of guessing where the cost leaks, this command walks through each lever and shows you where to cut.Source →

Claude Code 2.1.247 The /claude-api skill now covers the Admin API

The /claude-api skill was updated to cover the Admin API, organization members, invites, workspaces, API keys, rate limit reports, workload identity federation, and CMEK.

What changes You can check and adjust your team's API keys or usage limits straight from inside Claude Code, without hopping over to the web console.Source →

Gemini Gemini Live lets you handle to-dos by voice

Google added a major set of productivity features to Gemini Live so users can delegate to-dos using just their voice.

What changes When you're on the move or your hands are busy, you can add to or work through your to-do list just by talking, without looking at a screen.Source →

Claude Code 2.1.246 Startup warning for Bash allow rules with a wildcard before the subcommand

Bash allow rules with a wildcard placed before the subcommand, like Bash(git * main), can also match options inserted before the subcommand. Claude Code now warns about this at startup.

What changes You'll now catch, right when a session starts, an allow rule that's broader than you intended and could let through commands you didn't mean to permit.Source →

Claude Code 2.1.246 New Auto mode tab in /permissions

The /permissions screen now has an Auto mode tab where you can view and edit the classifier rules that decide what Auto mode allows automatically.

What changes You can see right in the app why Auto mode allows one command and blocks another, and adjust the rule on the spot if it's not what you want.Source →

Claude Code 2.1.246 Fixed leftover entries from crashed headless sessions on Windows/macOS

Fixed a Windows/macOS bug where stale entries left in ~/.claude/sessions by headless sessions that exited uncleanly weren't being cleaned up.

What changes If you run headless sessions repeatedly, the session list won't keep filling up with leftover entries from ones that crashed.Source →

Claude Code 2.1.246 /cd now applies the new directory's settings immediately

After moving to a new directory with /cd, that project's settings, hooks, MCP servers (still behind the usual approval prompt), skills, and agents now take effect immediately, without needing --resume.

What changes When you hop between projects, that project's setup is already active without having to restart the session.Source →

Copilot PowerPoint Copilot in PowerPoint can now pull content straight from an email

When asking Copilot in PowerPoint to build a presentation, you can now point it to a specific email as source material, instead of manually copying the content over.

What changes When you need to turn a report that arrived by email into a slide deck, this cuts out the manual copy-paste step.Source →

Copilot PowerPoint In PowerPoint Live, attendees can select slide text and have Copilot explain it

In PowerPoint Live meetings, attendees can now select text on a slide and get an instant explanation from Copilot, whereas slide content used to be static and non-interactive.

What changes If something on a slide isn't clear during a meeting, you can get an explanation on the spot without interrupting the presenter.Source →

Copilot Word Word's Read Aloud now answers spoken questions in real time

While Word reads a document aloud, you can now ask questions like 'What does this mean?' or 'Summarize this section' out loud and get an instant answer from Copilot, instead of pausing to search manually.

What changes If you're listening to a document while commuting or otherwise hands-free and something's unclear, you can ask about it out loud without stopping playback.Source →

Copilot Word Word's Copilot model picker now includes Claude Sonnet 5

Claude Sonnet 5 is now available in Word's Copilot model menu, using a higher default reasoning level for more accurate drafting and edits.

What changes When polishing a complex report or proposal, switching the model to Sonnet 5 is likely to produce a more accurate draft than before.Source →

Copilot Microsoft 365 Copilot app Generate and edit images right inside Copilot Cowork

Copilot Cowork now integrates OpenAI's ChatGPT Images 2.0 model, letting you generate and edit images without leaving Cowork for an external tool.

What changes You no longer need to switch to a separate image tool while drafting a report or deck, you can generate the image right next to the document you're working on.Source →

Copilot Microsoft 365 Copilot Chat Pick your own model for the Researcher agent in Copilot Chat

When adding the Researcher agent in Copilot Chat, users can now choose which model handles the research; previously users had no control over which model was used.

What changes You can pair a fast model with quick lookups and a more thorough one with research that needs solid sourcing, matching depth and speed to the task at hand.Source →

Copilot OneNote Bring Teams meeting content straight into Copilot Notebooks

Copilot Notebooks can now pull in a Teams meeting's transcript, notes, chat, and shared content as reference material, linking meeting context directly into a project's knowledge base to improve outputs like summaries or presentations.

What changes When wrapping up a project that spanned several meetings, you can gather all those meeting records in one notebook instead of hunting down each transcript separately, then generate a summary right there.Source →

Copilot Outlook Schedule meetings with Copilot chat in classic Outlook for Windows

In classic Outlook for Windows, users can now schedule meetings directly by chatting with Copilot or clicking the "Schedule with Copilot" button, which automatically checks attendee availability, books a room, drafts an agenda, and sends invites.

What changes Instead of trading several emails to find a time that works for everyone, you can hand the whole thing to Copilot in one request.Source →

Copilot PowerPoint PowerPoint Copilot can now pull slide content straight from an email

When asking Copilot in PowerPoint to build a presentation, users can now point to a specific email by subject, sender, or date and have its content pulled directly into the slides, with no manual copy-pasting required.

What changes When you need numbers or decisions from an email in your deck, you can just point Copilot at that email instead of copying and pasting the content yourself.Source →

Copilot Microsoft 365 Copilot Copilot now grounds answers in private Viva Engage content

Microsoft 365 Copilot now draws on private Viva Engage community and event content when answering questions, showing only what each user already has permission to see.

What changes Even discussions that only happened in a private internal community now surface when you ask Copilot, so you don't have to dig through that community yourself.Source →

Copilot Microsoft 365 Copilot Bring Outlook emails straight into Copilot Notebooks

Copilot Notebooks can now pull in Outlook emails as reference material directly, bringing the conversations, decisions, and context from those emails into the notebook.

What changes You no longer have to copy decisions out of an email thread by hand, you can add the email itself as a knowledge source and have Copilot summarize or organize straight from it.Source →

Copilot Microsoft 365 Copilot Copilot unifies web and work chat behind one Work IQ button

Microsoft merged Copilot's separate web chat and work chat into one interface, with a new Work IQ button in the top-left corner to toggle access to company data on or off.

What changes You no longer have to switch screens to move between general search and company-data search. Microsoft is tackling the same chat-versus-work-mode problem TechCrunch called out industry-wide, the same direction Anthropic and others are moving.Source →

Copilot Microsoft 365 Copilot Steer and auto-create Copilot pages on mobile

The Android and iOS Copilot apps now let users refine documents with commands like "shorten this page" or have Copilot auto-create a new page straight from a chat request.

What changes You can trim a report draft or spin up a new document straight from your phone while on the move, without needing to sit down at a computer.Source →

Copilot Microsoft 365 Copilot Microsoft 365 Copilot app gets a simplified design

The Microsoft 365 Copilot app got a design refresh across Windows, Mac, and web, simplifying the chat home screen, response layout, and navigation pane.

What changes With fewer menu options to parse, you spend less time figuring out where to click.Source →

Copilot Excel Excel Copilot can now edit with Python

Excel's Edit with Copilot feature can now run Python code, putting the results of statistics, simulations, and data transformations directly into your worksheet.

What changes You no longer need to open a separate Python environment for complex stats or simulations, just ask Copilot inside Excel and get the result in your worksheet.Source →

Copilot Microsoft 365 Copilot Copilot Notebooks gets a streamlined UI

Copilot Notebooks has a cleaner layout, simplifying how you open items and navigate the product.

What changes Less time spent hunting for the right item when you're juggling multiple notebooks.Source →

Copilot Microsoft 365 Copilot Copilot Chat now available while using Copilot Search

While using Copilot Search inside the Microsoft 365 Copilot app, you can now open Copilot Chat in a side pane at the same time.

What changes You can ask follow-up questions right next to your search results instead of switching pages and losing your context.Source →

Copilot Microsoft 365 Copilot Open Outlook emails directly inside Copilot Chat

Clicking an email that Copilot references now opens it right alongside the chat, instead of switching to a separate Outlook window.

What changes When Copilot cites an email as its source, you can check the original without switching windows.Source →

Claude app Claude's memory now spans chat and Cowork

Claude's memory now works across both chat and Cowork in the cloud. Everything it remembers appears under Topics in Settings, where you can edit or delete items, and sensitive topics like health or beliefs stay out unless you turn that on.

What changes You no longer have to re-explain in Cowork what you already told Claude in chat, cutting out the repeated context-setting when you move between tasks.Source →

ChatGPT ChatGPT Work can now complete tasks on sign-in-required sites

ChatGPT Work's browser can now complete tasks on some websites that require signing in. When a login screen appears, you enter your credentials yourself, and ChatGPT takes over the task from there.

What changes For a task on a system that requires login, like pulling a report or filling out a form, you can log in yourself and then step away while ChatGPT finishes the rest.Source →

Copilot PowerPoint Copilot in PowerPoint Can Now Cite Web Sources

When you ask Copilot in PowerPoint to build a presentation, it can now pull in and cite relevant web sources.

What changes You spend less time hunting down supporting sources yourself, so slides with citations come together faster.Source →

Copilot Word Copilot in Word Now Lets You Pick Anthropic Models

When editing a Word document with Copilot, you can now choose an Anthropic model in addition to OpenAI's.

What changes If one model's output doesn't land, you can switch to a different provider's model right away.Source →

Claude Code 2.1.239 Claude Code's Image Paste and Clipboard Tools Now Work on Alpine Linux

On musl-based Alpine Linux builds, the image-paste, clipboard, and audio-capture add-ons, previously built only for glibc and refused by the runtime, now load correctly.

What changes If you work inside Alpine-based containers, you no longer need a workaround just to get these add-ons running.Source →

Claude Code 2.1.239 The Usage-Limit Message Now Shows When Your Limit Resets

The message shown when you've used up your monthly spend limit now also states when your session or weekly limit resets.

What changes You can tell when you'll be unblocked without digging through docs, since the message tells you directly.Source →

Claude Code 2.1.239 A Post-Compaction Reminder No Longer Misfires a Skill's Original Arguments

The reminder shown after a session compacts was improved so a skill's original arguments no longer get re-run as if they were a brand-new request.

What changes Even after a long session compacts, a skill you already ran is less likely to fire again on its own.Source →

Claude Code 2.1.239 Long File Paths Now Truncate in the Middle to Fit One Line

When a file path shown on a tool-use row is long, it now truncates in the middle instead of wrapping or overflowing, keeping it on one line.

What changes Both ends of the path, the parent folder and the filename, stay visible, so you can tell which file it is without the row wrapping.Source →

Claude Code 2.1.239 Remote Sessions No Longer Time Out During Long Startup Hooks

Remote sessions now keep sending keep-alives while a long SessionStart or Setup hook runs, so the container isn't mistakenly reaped as idle mid-hook.

What changes Even with a heavy install script wired into a hook, you're less likely to get disconnected and have to start the session over before it finishes.Source →

Grok Grok Bot arrives

A persistent AI teammate with its own cloud machine: you message it, approve its actions, connect it to other services, and set up routines.

What changes Closer to handing work off and reviewing approvals than sitting with it through every step.Source →

Claude Code 2.1.239 Cost figures now include the US-only inference premium

In data-residency workspaces, /cost, the status line and --max-budget-usd all count the 1.1x premium.

What changes Budget caps stop being quietly optimistic.Source →

Claude Code 2.1.239 Bedrock, Vertex and Foundry get the fullscreen offer too

The one-time fullscreen renderer prompt now reaches setups that were previously excluded, and fresh installs there start fullscreen.

What changes People going through a cloud provider on a company account start from the same screen as everyone else.Source →

Claude Code 2.1.239 A command that moves your Python project to the 1.x SDK

/claude-api upgrade migrates anthropic 0.x projects to 1.x, and the skill's Python reference now targets 1.x. Timeouts use anthropic.Timeout, not httpx.Timeout.

What changes It sweeps the places a hand migration tends to miss.Source →

Claude Code 2.1.239 Synced plugins are labelled separately in cloud sessions

Plugins synced from claude.ai show as name@synced and toggle with claude plugin enable/disable <name>@synced. They never override a same-named plugin you installed yourself.

What changes No more wondering which copy of a plugin is actually running.Source →

ChatGPT Plugin recommendations now favour what people keep using

Rankings put plugins people still open after installing at the top. Web and mobile for now, not the desktop app.

What changes Less chance a genuinely useful plugin sits buried under ones nobody opens twice.Source →

ChatGPT Interactive content shows up while it is still being built

It now appears progressively in the conversation instead of arriving all at once when finished.

What changes Less staring at nothing, and you can tell it is heading the wrong way before it gets there.Source →

ChatGPT Generated images fill the screen on Android

Images now span the full width of the conversation, and the sidebar lists up to eight recent chats. App version 1.2026.216 and later.

What changes You can judge an image without tapping to enlarge it, and yesterday's thread is easier to get back to.Source →

ChatGPT ChatGPT now knows what time it is for you

It reads your local time more accurately during a conversation.

What changes You can say "by tomorrow morning" and mean it. Before, a timezone gap could land it on the wrong day.Source →

ChatGPT Long conversations open faster on the web

Messages load in chunks instead of pulling the whole thread at once.

What changes Less waiting when you reopen a project thread you have been running for weeks.Source →

ChatGPT Recent photos are one press away on iPhone

Press and hold the + button and your recent photos come up.

What changes One less step when you photograph a whiteboard or a document to ask about it.Source →

Codex rust-v0.149.0 · New Features Codex gets a task dashboard

An interactive view for searching, starting, opening, renaming and stopping tasks, with shortcuts you configure.

What changes When several sessions run at once, less time hunting for which window was doing what.Source →

Codex rust-v0.149.0 · New Features You can change directories inside a session

New /cd, /pwd and /cwd commands for TUI sessions.

What changes Moving between repos no longer means killing the session and starting over.Source →

Codex rust-v0.149.0 · New Features Queue your next instruction into a running session

codex queue sends messages to existing local or remote sessions.

What changes Instead of waiting for a task to finish before typing the next step, you queue it and go do something else.Source →

Codex rust-v0.149.0 · New Features Vim editing covers more ground

Character replacement plus change motions like cw, c$ and cc.

What changes Long prompt edits go at your usual speed instead of breaking your hands out of habit.Source →

Codex rust-v0.149.0 · New Features codex doctor says more about why it will not connect

It now checks endpoint protection, network and proxy failures, desktop app state and update connectivity.

What changes When it will not run on a locked-down work laptop, there is somewhere to look before you start guessing.Source →

ChatGPT Quizzes and dictation land in the ChatGPT app

It can turn a conversation into a quiz, and you can move between typing and voice dictation.

What changes Useful for checking whether you actually absorbed a document, and for drafting out loud while you are moving.Source →

ChatGPT You can change memory settings on a project you already made

Open the project, use the three-dot menu, and switch how it remembers.

What changes You can turn memory off for the project holding company material while leaving it on elsewhere.Source →

ChatGPT Your Drive files open inside ChatGPT

With the Drive plugin connected, Library browses your files and folders and @-mentions pull one into a chat. Docs, Sheets and Slides can sit open beside the conversation, and you can point it at a whole folder.

What changes No re-uploading the same file. The original stays in Drive, so you can edit it and ask again.Source →

ChatGPT The Mac app can remember what you were working on, off by default

It logs interaction events from apps and sites you pick so ChatGPT and Codex can refer back. No screenshots, recordings or microphone input, and private browsing is excluded. Pro, Business and Enterprise, not yet in the EEA, UK or Switzerland.

What changes You stop re-explaining context every time. It is a recording feature, so you choose which apps count and can inspect or delete what it kept.Source →

Copilot Microsoft 365 Copilot Planner Agent reaches the basic plans

It used to be premium only. Now every group-based Planner plan gets it, as long as you hold a Microsoft 365 Copilot licence.

What changes You no longer upgrade the plan just to hand it tasks and have it report back.Source →

Copilot Outlook Outlook comments on your wording while you write

Copilot gives feedback on tone and structure in a side chat as you draft.

What changes Worth having when there is nobody to read it over first, especially for a second-language email or an awkward one.Source →

Copilot Outlook Outlook pulls together what you need before a meeting

Copilot gathers the mail, files and past discussion tied to your next meeting.

What changes On a day of back-to-back meetings, less time spent remembering what the last one was about.Source →

Copilot PowerPoint Copilot builds slides in web PowerPoint too

What used to need the installed app now works in the browser.

What changes You can put a first draft together from a machine that is not your work PC.Source →

ChatGPT ChatGPT Can Now Book Restaurant Reservations

ChatGPT now connects to OpenTable, Resy, and Yelp, so telling it where, when, party size, and vibe surfaces open reservation slots directly in the chat.

What changes You can lock in a dinner reservation in one conversation instead of jumping between separate booking apps.Source →

ChatGPT ChatGPT Voice Now Supports Files and Projects

You can now upload a file mid voice conversation and ask about it, or use voice inside a Project referencing its recent chats and sources.

What changes Even when your hands are busy, you can keep working with your documents just by talking instead of typing.Source →

Claude app Claude app adds automatic malware scanning for third-party skills and plugins (beta)

Enterprise plans can now turn on skill and plugin security scanning, which automatically checks third-party skills and plugins for malicious content whenever someone uploads or edits them.

What changes When a team member wants to install a skill or plugin whose origin is hard to verify, an admin who has turned this on gets malicious code filtered out automatically.Source →

ChatGPT GPT-5.6 Gets More Accurate, and Free Users Get Unlimited Chats

Plus and Pro users get an updated GPT-5.6 Sol with more reliable facts and a slider to control how much thought goes into each answer. Free and Go users get GPT-5.6 Luna as their default model this week, followed by unlimited text chats and a new Think button next week.

What changes Both answer quality and how much you can chat improve regardless of your plan, which quietly narrows one reason free users had to upgrade.Source →

ChatGPT Long Pastes Now Turn Into Attachments Automatically

Paste more than 10,000 characters into the composer and ChatGPT now converts it into an attachment instead of filling the text box with raw text.

What changes Dropping in a long log or document no longer clutters the composer, so you can keep the conversation going right after.Source →

ChatGPT ChatGPT Work and Codex Get Plugins for Students and Educators

Three new plugins let students build study guides, quizzes, and flashcards from their own course materials, and let educators draft syllabi and course content.

What changes Instead of re-explaining your context every time, you get a tool already set up for studying or lesson prep.Source →

ChatGPT The DALL·E GPT Inside ChatGPT Is Being Retired on August 30

The official DALL·E GPT built into ChatGPT goes away on August 30. Image generation itself continues through ChatGPT Images, and user-created GPTs with image generation aren't affected.

What changes If you want to keep images you made with it, download them before August 30, since they won't be reachable through it afterward.Source →

Copilot Microsoft 365 Copilot Chat Microsoft 365 Copilot can now list your email attachments in chat

Microsoft 365 Copilot Chat can now list the file attachments you've sent or received over email, without you having to search your inbox manually.

What changes You can just ask, 'what attachments did I get last week,' instead of digging through your inbox.Source →

Copilot Microsoft 365 Copilot Chat Take a screenshot directly inside Copilot

Windows Copilot now has a built-in screenshot tool, so you can capture your screen and attach it to a prompt without leaving Copilot.

What changes You can capture an error or a table on screen and ask Copilot about it on the spot, without a separate screenshot tool.Source →

Copilot Microsoft 365 Copilot Chat Copilot Chat Can Now Pull In SharePoint Lists Directly

In Copilot Chat, you can now search for and select SharePoint Lists directly through the Context IQ menu while composing a prompt. Previously there was no way to reach Lists from that menu.

What changes If your team tracks things like inventory, schedules, or requests in a SharePoint List, you can now pull that data straight into the chat instead of copying it in by hand.Source →

Copilot Copilot Notebooks' Overview Screen Gets a Cleaner Redesign in OneNote

The Copilot Notebooks Overview experience in OneNote has been redesigned, offering crisper AI-generated summaries and key insights, plus one-click artifacts to keep work moving.

What changes Teams that collect meeting notes in OneNote can now scan summaries and key points faster from the redesigned overview screen.Source →

Copilot Microsoft 365 Copilot Copilot responses now show images from your files and meetings inline

Copilot responses now surface relevant images from your files and meetings directly inline, instead of returning text only as before.

What changes When you ask about material that includes a chart or diagram, you can see the actual image alongside the explanation instead of just reading about it.Source →

ChatGPT ChatGPT launches a free workspace for academic researchers

OpenAI launched ChatGPT for Academic Researchers, letting verified small teams of faculty and postdocs apply for 12 months of free access to a dedicated ChatGPT workspace, supporting up to five members with business-grade data protection.

What changes A department or individual researcher without much funding can now apply and get a year of paid-tier ChatGPT for free.Source →

ChatGPT Sign in with ChatGPT rolls out

On select partner sites, including Airtable, GitLab, HubSpot, Notion, Supabase, and Vercel, you can now use Sign in with ChatGPT to create or link an account.

What changes You can sign up for these services using your existing ChatGPT login instead of creating a new password each time.Source →

Gemini Gemini on Mac now edits by voice, not just transcribes

You can speak instructions and get corrections and summaries back.

What changes Worth trying when your hands are busy, or when a long meeting recording needs to become notes.Source →

Claude app Claude Opus 5 is out

Anthropic describes it as thoughtful and proactive, close to Fable 5's intelligence at half the price.

What changes Fewer jobs need the most expensive tier. Long documents and multi-step work are where the difference shows.Source →

ChatGPT ChatGPT Voice now works inside Work and Codex

ChatGPT Voice is now available inside Work and Codex on the desktop app, letting you start a task by speaking naturally and interrupt or redirect it mid-stream.

What changes You can direct work and check progress just by talking instead of typing, so you can hand off a task while your hands are busy with something else.Source →

ChatGPT Health data now shows up inside ChatGPT

Logged-in US users 18 and older can now connect supported health records and Apple Health data to see a dashboard of lab results and medications, and ask ChatGPT questions grounded in that personal health context.

What changes When you get lab results and aren't sure what they mean, you can ask ChatGPT directly instead of digging through the report yourself.Source →

ChatGPT ChatGPT desktop app makes it easier to switch modes and continue work

The ChatGPT desktop app was updated to make it easier to switch between Chat and Work, find past conversations and Projects, and continue Work across devices, now live for all plans on macOS and Windows.

What changes You can pick up a task on your laptop right where you left off at your desk, without re-explaining the context every time you switch machines.Source →

ChatGPT ChatGPT desktop app gets easier navigation and more voice

The ChatGPT desktop app makes it easier to switch between Chat and Work, find conversations and Projects, and continue Work across devices. Voice has also expanded to Work and Codex, so you can start a task by speaking and interrupt it mid-task.

What changes Picking up a task when you switch between your laptop and desktop gets easier, and when your hands are busy, you can hand off a task by talking instead of typing.Source →

ChatGPT ChatGPT raises the custom instructions character limit

The character limit for ChatGPT custom instructions increased from 1,500 to 5,000 for Plus, Pro, Enterprise, Business, and Education users.

What changes You can now spell out your company's style guide or recurring requirements in far more detail, instead of re-explaining the same conditions every time.Source →

ChatGPT ChatGPT search now covers chats, projects, and files at once

ChatGPT rolled out a new way to search chats, projects, images, and documents in one place on web, iOS, and Android, searchable straight from the sidebar with filters to narrow results.

What changes Even if you can't remember which conversation you mentioned something in weeks ago, one search now finds it across chats and documents alike.Source →

ChatGPT ChatGPT returns to WhatsApp in the European Economic Area

ChatGPT is available again on WhatsApp in the European Economic Area, reachable by messaging a designated number without needing a ChatGPT account, supporting chat, image generation, and voice notes.

What changes European users can now call on ChatGPT straight from the WhatsApp app they already use, without installing anything new.Source →

Claude app Claude's memory is now a list of entries

What used to be lumped together by day is stored as separate topics and updated during the conversation.

What changes You can see what it holds as discrete items, so a wrong one is easy to find and delete.Source →

ChatGPT ChatGPT Work is a new agent that carries tasks through to the end

ChatGPT now has Work, an agent for longer, more involved tasks. It researches and analyzes across connected apps and files and produces finished documents, spreadsheets, presentations, reports, and Sites, while you watch its progress, redirect it, or approve key actions. It can also run on a schedule, once, repeating, or triggered by changes.

What changes For tasks that used to mean digging through multiple sources to draft a report, you can now hand the legwork to Work and just steer it along the way.Source →

ChatGPT ChatGPT Skills is now generally available

ChatGPT Skills is now generally available. Enterprise and Edu users can find it under Plugins in the sidebar, on the Skills tab.

What changes Once you register a repeated workflow as a skill, ChatGPT can follow that procedure without you retyping the same instructions every time.Source →

ChatGPT The new ChatGPT desktop app brings Chat, Work, and Codex together

The new ChatGPT desktop app is available globally for macOS and Windows, combining Chat for conversation, Work for research and finished deliverables, and Codex for software development in one app. With permission, Work can use local files and desktop apps, and a built-in browser lets ChatGPT gather information from the web.

What changes If you've been switching between separate windows for chatting, drafting documents, and coding, you can now do all three from one app.Source →

ChatGPT ChatGPT Sites lets you build a website without leaving the chat, now in public beta

ChatGPT Sites is now in public beta. Without leaving ChatGPT, you can turn work or ideas into a website or lightweight app, dashboards, project trackers, launch calendars, prototypes, internal portals, reports, previewing and refining the result before sharing. Business and Enterprise customers already had access; now they can also publish Sites publicly and share them via a URL with anyone.

What changes When you need a simple status page or internal portal, you can now build it through a ChatGPT conversation, no separate web development, and hand it off as a single link.Source →

ChatGPT ChatGPT is retiring group chats

Starting July 9, users on web, iOS, and Android can no longer create new group chats, convert existing conversations into group chats, or join one through an invite link. Existing group chats remain usable, and even once one turns read-only, its messages, files, and images stay accessible. Individual conversations are unaffected.

What changes If you've been using a group chat with colleagues, it still works for now, but since you can't create new ones, you'll need another way to start fresh group collaboration.Source →

Claude app It shows where your month actually went

Settings > Reflect now has a monthly recap: topics covered, busiest day, peak hours.

What changes Something to point at when you write a retrospective or a status report and need to recall what you were really buried in.Source →

ChatGPT ChatGPT Voice switches to a new model that handles overlapping speech naturally

ChatGPT Voice now runs on new models, GPT-Live-1 for paid users and GPT-Live-1 mini for Free users. Both can listen and speak at the same time, making turn-taking and interruptions feel more natural. It works inside a ChatGPT chat, showing streamed text alongside spoken responses, and can use web search, memory, and supported widgets to show visual results.

What changes You can interrupt mid-sentence with a new question and it responds naturally, making voice chats feel more like an actual phone call.Source →

Claude app Claude Cowork moves to web and mobile

Desktop-only until now. Rolling out from the Max plan onward.

What changes You can check progress and give the next instruction without being at your desk.Source →

Claude app Claude can act on your Outlook mail and calendar

The Microsoft 365 connector goes beyond search: drafting and sending mail, creating calendar events.

What changes "Reply with this and book a follow-up next Tuesday" becomes one step. It is write access, so decide deliberately what you allow.Source →

ChatGPT ChatGPT Now Builds PowerPoint Slides From Inside the App

ChatGPT for PowerPoint is now generally available worldwide. From a sidebar inside PowerPoint, you can create new slides and decks, revise existing presentations, and ask questions about a deck's story and structure.

What changes You no longer need to switch between a ChatGPT tab and PowerPoint to build a deck. When you're up against a deadline, you can ask about slide structure right from the sidebar.Source →

Claude app Access to Claude Fable 5 and Claude Mythos 5 Is Restored

Access to Claude Fable 5 and Claude Mythos 5, suspended for a while, has been restored.

What changes If you had switched workflows to another model while these were down, you can move back now.Source →

Claude app Claude Sonnet 5 is out

Anthropic's most agentic Sonnet yet, with gains over Sonnet 4.6 in reasoning, tool use, coding and knowledge work.

What changes The mid tier now covers a wider range of work before you need to move up.Source →

ChatGPT ChatGPT's Personal Finance Feature Reaches Plus Users and Android

ChatGPT's personal finance feature is rolling out to Plus users in the US on web and iOS, and is now available on Android for Pro and Plus users. Eligible users can connect financial accounts and view a dashboard while asking ChatGPT questions.

What changes Instead of opening several banking apps on payday, you can now ask ChatGPT directly about this month's spending in one place.Source →

ChatGPT ChatGPT Dictation Gets More Accurate

ChatGPT rolled out a new speech-to-text model for dictation across all plans. There's no change to how you use it, but transcription accuracy improved across languages, accents, and when speakers mix languages mid-sentence.

What changes If you dictate notes on the go or mix English and Korean mid-sentence, you'll retype fewer corrections afterward.Source →

ChatGPT Codex Remote Goes GA, Letting You Approve Coding Sessions From Your Phone

Codex Remote is now generally available on all ChatGPT plans. From the ChatGPT mobile app, you can start or continue work on a connected Mac or Windows host, check progress, and approve actions. Remote Control now uses authenticated one-to-one QR pairing per device.

What changes You can kick off a long coding task and step away, then check progress and approve the next step from your phone even in a meeting.Source →

ChatGPT ChatGPT's Default Model, GPT-5.5 Instant, Gets Better at Advice and Planning

OpenAI updated GPT-5.5 Instant, ChatGPT's most-used model, to improve conversational quality, especially when users are making a decision, asking for advice, planning something, researching options, or shopping. The model better identifies the underlying goal behind a question.

What changes When you're working through a multi-step decision like a job change or a move, the conversation flows more naturally without having to re-explain context each time.Source →

Claude app You Can Now Delegate Work to Claude Directly in Slack

On Team and Enterprise plans, you can now tag Claude in a Slack conversation and hand off a task while you do something else.

What changes You don't need to switch over to a separate Claude app to delegate something, since the Slack thread you're already in is enough.Source →

ChatGPT ChatGPT Now Turns Long Pastes Into Attachments Automatically

For ChatGPT Free and Go users, pasting more than 10,000 characters into the composer now automatically converts the content into an attachment instead of inserting it directly into the text field.

What changes Pasting a full meeting transcript or long report no longer clutters the composer, and you can refer back to it as a separate attachment instead of scrolling through the chat.Source →

ChatGPT Codex's New Record & Replay Turns a Demo Into a Reusable Skill

The Codex app for macOS added Record & Replay: demonstrate a repetitive workflow once, and it turns that demonstration into a reusable skill you can run again automatically.

What changes If you have a repetitive task that's easier to show than describe, like filing a configured issue or downloading a recurring report, this lets you automate the process itself. It requires Computer Use to be enabled, and initial availability is limited.Source →

ChatGPT ChatGPT's Scheduled Tasks Get Their Own Management Page

ChatGPT added a dedicated Scheduled page, accessible from the sidebar, where you can see all active tasks, check when they'll run next, and pause, resume, edit, or delete them in one place.

What changes If you've been having ChatGPT handle reminders or recurring work, you no longer have to dig through past chats to find them, everything's managed from one screen.Source →

ChatGPT GPT-5.2 Models Retired, Conversations Auto-Switch to GPT-5.5

As of June 12, 2026, GPT-5.2 Instant, Thinking, and Pro are no longer available in ChatGPT. Existing conversations that used GPT-5.2 automatically continue on the corresponding GPT-5.5 model.

What changes If you had a workflow that specifically called GPT-5.2, your conversations keep working without any action needed, but know that GPT-5.5 is now the one actually responding.Source →

ChatGPT ChatGPT Adds Finer Controls Over Memory

You can now delete individual memories from the memory summary page, and turn memory off entirely via "Delete and turn off memory" in the three-dot menu. This doesn't delete your past chats.

What changes If a specific thing ChatGPT remembers bothers you, you can now delete just that item instead of turning memory off entirely. Note that if you turn memory back on later, it may recreate memories from chats still in your history.Source →

Claude app Access to Claude Fable 5 and Claude Mythos 5 Has Been Suspended

Anthropic suspended access to Claude Fable 5 and Claude Mythos 5, saying it aims to restore access as quickly as possible, with details in its official statement.

What changes If you were relying on either model for work, you'll need to switch to another Claude model for now. The most reliable way to check restoration status is Anthropic's own notice or announcement.Source →

Claude app You Can Now Edit Claude's Drafts in Place

When Claude writes a draft, like a report or a plan, in chat or Claude Cowork, it opens right beside the conversation. Highlight the part you want changed, type the change, and it's edited right there.

What changes You don't have to re-describe a whole paragraph or switch apps, just point at the part you want changed.Source →

Claude app Claude Fable 5 launched

Anthropic launched Claude Fable 5, a Mythos-class model made safe for general use.

What changes Capabilities from Anthropic's top-tier model family, previously aimed mainly at enterprise customers, are now available to everyday users through Fable 5.Source →

Claude app Claude Opus upgraded to version 4.8

Claude Opus has been upgraded to version 4.8. Anthropic says it improves on the previous 4.7 version in coding, agentic tasks, reasoning, and practical knowledge work.

What changes If you've been using Opus for complex reports or multi-step task automation, the same requests are now more likely to come back better.Source →

Notion Simple tables can merge cells

Cell merging now works in plain tables, not only databases.

What changes Headers can span columns and related rows can be grouped, so the explanation stays inside the table.Source →

Notion Notion adds a planning mode

A screen for laying out what happens in what order before you start.

What changes When several people are involved, you can agree on sequence before anyone makes a page.Source →

Notion Notion's mobile home screen is rebuilt

The layout you land on when you open the mobile app has been reworked.

What changes Fewer taps to reach the pages you open from your phone all the time.Source →

Claude app Anthropic Launches Claude Design for Collaborative Visual Work

Alongside Opus 4.7, Anthropic Labs launched Claude Design, a new product that lets you collaborate with Claude to create visual outputs like design mockups, prototypes, slides, and one-pagers.

What changes Instead of switching to a separate design tool, you can now create and refine visual materials, like a deck or a mockup, directly through conversation with Claude.Source →

Grok Grok retires the grok-imagine-image-quality model on November 2

Source says September 2026 · checked 2026-09-08

Starting November 2, requests to grok-imagine-image-quality will be served by grok-imagine-image-2.0 with quality set to low, at no change to the request or response format and a lower per-image price. grok-imagine-image (1.0) is unaffected.

What changes If your code or automation calls this model by name, requests auto-route to the replacement, but output quality may shift, so compare actual results before November 2.Source →

Grok Grok Imagine's video model adds text-to-video at 1080p

Source says July 2026 · checked 2026-09-08

grok-imagine-video-1.5 now supports text-to-video, image-to-video, and reference-to-video, with optional preset voices, and native 1080p for the text- and image-based modes.

What changes You can generate a 1080p video straight from a text description with no source image needed, skipping the shoot-or-source-image step for short promo or social clips.Source →

Perplexity Perplexity's Pro Search reaches general availability

Source says November 2025 · checked 2026-09-02

Pro Search, which breaks a question into multiple automated web searches and reasoning steps, is now generally available on the Sonar Pro model.

What changes You can ask a complex research question in one go and Perplexity will break it into multiple searches on its own.Source →

Perplexity Attach files directly to your Perplexity search

Source says September 2025 · checked 2026-09-02

You can now upload PDF, Word, and text files and ask questions about them or get a summary.

What changes You can upload a report or contract as-is and ask for just the key points.Source →

Perplexity Image uploads now open to all Perplexity users

Source says April 2025 · checked 2026-09-02

Uploading an image to use as part of a search is now available to every user, not just paid tiers.

What changes Snap a photo and ask about what's in it right away.Source →

Grok Grok's voice mode upgrades to Think Fast 2.0

Source says July 2026 · checked 2026-09-02

Grok's voice conversation model is upgrading to Think Fast 2.0, with the existing voice mode automatically routing to it starting August 5.

What changes No settings to change, your next voice conversation should already feel faster.Source →

Grok Grok's web search adds image search

Source says May 2026 · checked 2026-09-02

Grok's web search can now explicitly search for images and show them in its replies.

What changes When something's hard to explain in words, you can get related images back alongside the answer.Source →

Grok Grok adds custom, cloneable voices

Source says May 2026 · checked 2026-09-02

You can now clone a voice from a short audio clip and use it for both text-to-speech and voice conversations.

What changes You can hear Grok's answers in your own voice or one you choose.Source →

Perplexity Perplexity Cuts GPT-5.6 Prices and Adds a Fast Mode

Source says July 2026 · checked 2026-09-01

Perplexity cut GPT-5.6 Luna's price to $0.20 per million input tokens and $1.20 per million output tokens, and GPT-5.6 Terra to $2 and $12 respectively. GPT-5.6 Sol now supports a Fast mode at 2x standard pricing for quicker responses.

What changes If your work uses Perplexity's API, you can pick the pricier Fast mode for urgent requests and save on the rest with the lower base pricing.Source →