Jiberish

AI News

The AI defaults, prices, and shutdowns landing before December

A dated list of what changes in Claude, ChatGPT, Gemini, and Grok between November 16 and December 11, 2026, and the decision each change asks of the person who runs your AI tools.

Kyle Del FranciaUpdated 7 min read
In this piece

Most AI changes don't need a meeting. The ones below do, because each one turns something on for your whole organization, changes what you pay, or switches off something an integration may depend on, and each has a date. If you run the Claude, ChatGPT, or Google Workspace admin console, or someone built an integration on one of these APIs for you, this is the list to check before Thanksgiving.

We checked each source linked here on November 10, 2026. Dates are as the providers published them and can move, so confirm in your own admin console before acting. Our Bench pages hold the fuller picture for each lab.

The calendar

DateWhat changesWho it affects
Nov 16Claude's unified chat and Cowork experience turns on by default for existing Enterprise organizationsClaude Enterprise owners
Nov 16Fable 5.2 launch week ends for Pro and Team standard seatsClaude Pro and Team
Nov 18ChatGPT dots on by default for members, and billing startsChatGPT Enterprise workspace admins
Nov 19Claude Data turns on by default for EnterpriseClaude Enterprise owners
Nov 30Google Workspace skills start reaching Scheduled Release domains, on by defaultWorkspace admins
Nov 30Gemini 4 Carbon's introductory API price endsGemini API users
Nov 30Claude Sonnet 4.5 retires from the Claude APIClaude API integrations
Nov 30OpenAI shuts down reusable prompts, Agent Builder, and EvalsOpenAI API integrations
Dec 1Gemini 2.5 Pro and 2.5 Flash requests start routing to newer modelsGemini API integrations
Dec 1Older OpenAI image models are removedOpenAI API integrations
Dec 11GPT-5, o3, and related OpenAI models are removedOpenAI API integrations

Claude Enterprise: two defaults in four days

On November 16, existing Enterprise organizations that haven't chosen get the new unified chat and Cowork experience automatically. Owners can pick "Enable now" or "Keep disabled" before then, and a choice already made stays in place (Enterprise setup). Anthropic's suggested rollout starts with a security review and a pilot through a custom role. The same page describes Efficient mode, which limits members to Sonnet and Haiku in the new experience and starts sessions on Sonnet 5.5 at low effort, a reasonable way to control cost during a pilot. Organizations with zero data retention, the HIPAA-ready configuration, or customer-managed encryption keys can't use the new experience at all.

Decide: enable now for a pilot group, or keep disabled until your review is done. Not deciding means it turns on.

On November 19, Claude Data, the renamed Claude Dashboards, turns on by default for Enterprise. It needs a data connector such as a warehouse or CRM to do anything. Sheets and Whiteboards, released alongside it in beta, stay off by default on Enterprise and are on by default on Team. All three are under Organization settings, then Artifacts (announcement).

Decide: whether anyone should be querying company data through Claude yet, and which connectors they can use.

For Pro and Team standard seats, Fable 5.2 is included up to half of weekly limits only until November 16 at 11:59 pm PT (launch offers). After that it runs on usage credits. Our Fable 5.2 guide has a one-week test to run before the window closes.

ChatGPT Enterprise: dots switch on and start billing

OpenAI's admin guide says that starting November 18, members have access to dots by default unless the workspace opts out, and billing for enabled dots begins the same day. Eligible Enterprise workspaces choose flat-rate billing, based on the type of work a dot does, or consumption billing. Tasks a dot hands off to ChatGPT Work or Codex threads are charged separately at standard rates. To limit access to a group, keep the workspace default off and assign a custom role with the "Use dots" permission (dots admin guide).

Decide: opt out, or limit dots to a role, before November 18, and pick a billing model. A dot is an always-on agent with its own cloud computer, so treat it like a new hire with system access: name an owner, set what it can touch, and review its work. Dots stay on Trial on The Bench until we can verify the billed terms.

Google Workspace: skills arrive on by default

Skills in Workspace began rolling out to Rapid Release domains on November 9 and start reaching Scheduled Release domains on November 30, with up to 15 days before people see them. They are on by default for eligible Business and Enterprise editions, and admins can turn them off for the domain, an organizational unit, or a group in the Admin console. They work in Gmail, Docs, Slides, Drive, Chat, and Workspace Studio, not yet in Sheets or Meet, and support English prompts only (Workspace Updates).

Decide: whether skills go to everyone or a pilot group first, and who may share skills outside the organization.

For API integrations: models and tools that stop working

If someone built software on these APIs for you, ask them to check for these. Each is easy to fix ahead of time and confusing to debug after.

  • Claude Sonnet 4.5 (claude-sonnet-4-5-20250929) retires from the Claude API on November 30. Requests to a retired model fail. Anthropic's recommended replacement is Sonnet 5.5 (model deprecations).
  • OpenAI shuts down the v1/prompts endpoint and reusable prompt objects, Agent Builder, and the Evals dashboard and API on November 30. On December 1 it removes gpt-image-1-mini, gpt-image-1.5, and chatgpt-image-latest. On December 11 it removes gpt-5, gpt-5-mini, gpt-5-nano, gpt-5-pro, o3, and o3-pro, with GPT-5.6 models as the listed replacements (deprecations).
  • Gemini 2.5 Pro and 2.5 Flash were deprecated on November 2. From December 1, requests to gemini-2.5-pro route to gemini-3.8-flash and requests to gemini-2.5-flash route to gemini-3.5-flash-lite, billed at the new model's rates (Gemini API changelog). These calls won't fail, which is the risk: the model answering them changes without anyone touching the code. Test the new model on your real inputs before then.
  • Gemini 4 Carbon became generally available on November 10 at an introductory $2 / $10 per million input and output tokens. After November 30 the price becomes $4 / $20. Google itself says Argon "remains the ideal choice" for most software and knowledge work (announcement).

Decide: who owns the check, and by what date. A model ID search across your code and automation tools takes minutes.

Already in effect

Two changes landed earlier this month and are worth a look now.

Microsoft is adding fair-use limits to more Copilot Chat features for people without a paid Microsoft 365 Copilot license. The rollout began November 2 for organizations with 2,000 or fewer Microsoft 365 seats and November 9 for larger ones. Admins can't see or change the limits, Microsoft doesn't publish them, and in most organizations the notice includes an option to request a license (Partner Center announcements). Decide who needs a paid seat before someone gets blocked mid-task.

If your team uses Grok Bot, its November releases let a Bot finish a purchase or booking on its own after a person approves the cart and its total (Grok Bot changelog). That's a useful step, and a reason to agree on spending limits and who can approve before anyone gives a Bot a payment method.

How to use this list

Pick one owner for AI admin settings if you don't have one. Give them this table, the admin console access each item needs, and thirty minutes this week. For each row, the answer is one of three: leave the default, change it now, or put a date on the calendar to revisit. Write the answer down next to the row.

If you'd like a second pair of eyes on the settings, or help rolling a change out with training, see AI Advisory & Training.

Put this into practice

Bring the decision you’re working through.

A free discovery session is a chance to review your goal, the tools or website you have, and the help you need. We’ll identify a sensible next step and discuss scope before any project commitment.

Custom AI Systems & Agents →AI Advisory & Training →Book a free call