ChatGPT moved to GPT-6.
ChatGPT's Chat tab moved to GPT-6, and answers can now arrive as diagrams and working tools. Check what those tools assume. Also: monthly API credits for Claude Max and Team, a cheaper Haiku, and audio uploads in ChatGPT.
As of Oct. 8, 2026 · About a 4-minute read
View PDFDownload4 pages · 142 KBChatGPT answers can now come as tools.
On Oct. 7, OpenAI moved ChatGPT's Chat tab to GPT-6, whose answers can include diagrams, comparisons and calculators you use in the chat. The same day Anthropic shipped a cheaper small model, Haiku 5.5, and gave Max and Team plans monthly credits for its developer platform. GPT-5.5 leaves Codex on Oct. 14, as covered last week, and Anthropic's updated Usage Policy applies from Nov. 12.
OpenAI GPT-6 · Anthropic Haiku 5.5 · Gizmochina · Anthropic usage policy- Oct. 6Audio uploads arrive in ChatGPT
- Oct. 7GPT-6 reaches paid ChatGPT plans
- Oct. 8GPT-6 starts reaching Free and Go
- Oct. 14GPT-5.5 leaves Codex
- Nov. 12Anthropic's updated Usage Policy applies
Before Oct. 14: swap GPT-5.5 out of saved Codex settings, custom agents and scripts.
OpenAI release notes · Gizmochina · Anthropic usage policySix changes this week
-
Try it
GPT-6 now answers in the Chat tab, with interactive layouts
Plus, Pro, Business and Enterprise get GPT-6 Sol; Free and Go get GPT-6 Luna. Answers can mix text with diagrams, side-by-side comparisons and working tools such as calculators, and can start before ChatGPT has finished thinking. Enterprise access depends on admin settings. The Pro reasoning option stays on GPT-6 Astra without the new layouts, and Work and Codex models don't change.
OpenAI GPT-6 · OpenAI release notesA calculator inside an answer OpenAI release notesGPT-6 can answer with a working calculator.
-
Turn it on
Max and Team plans now include monthly API credits
Max 5x gets $100 a month, Max 20x gets $200, and Team gets up to $500 pooled across its users, for running your own apps and agents on the Claude Platform. Anthropic says they work on any of its models. To claim them, link a Claude Console organization in Settings > Billing (Max) or Organization settings > Billing (Team). On a work plan, claim and spend them only if your company allows it.
Claude release notes · Anthropic Haiku 5.5Claude's new monthly API creditsDollars a month, by plan For your own apps and agents on the Claude Platform. Claim them in billing settings.
Anthropic Haiku 5.5 · Claude release notes -
Try it
Haiku 5.5, a cheaper small model for the API
For prompts up to 100,000 tokens, Anthropic lists $0.10 per million input tokens and $0.50 per million output tokens, against $1.00 and $5.00 for Haiku 4.5, and says it costs around 75% less to run on average. It's the first Haiku with an adjustable effort setting. Anthropic says Sonnet 5.5 and Opus 5.5 remain better for complex agentic coding. Test it on your high-volume jobs before you switch.
Anthropic Haiku 5.5 -
Try it
Upload audio for transcripts, summaries and questions
Attach an audio file to a chat and ask for a transcript, a summary or a follow-up email draft. OpenAI says availability varies by workspace settings, region, app version and model, and that transcripts may contain errors and vary by language. Check names, numbers and decisions against the recording before you send anything on. If the recording includes other people, ask them first, and at work do it only if your company allows it.
OpenAI release notes -
Try it
Copilot Chat can redo an answer, or hand it to another model
At the end of a response, open More options and choose Try Again for a fresh answer, or Switch Model to get one from a different AI model, without retyping the prompt. The same release notes add chat inside Copilot Search on Windows and the web, and custom skills in Copilot in PowerPoint for Windows. Admins can now edit Copilot connector rules themselves, under Settings > Search & intelligence > Data sources. Microsoft rolls features out to some people in a company first.
Microsoft 365 Copilot release notes -
Heads-up
Anthropic's updated Usage Policy takes effect Nov. 12
Anthropic says most of the changes clarify existing rules. Where Claude's output can affect someone's health, legal rights, finances, livelihood or access to essential services, it still requires a qualified human in the loop, and the person affected must be told AI was used. The update also spells out rules against networks of fake accounts and fabricated news sites. If your team uses Claude where those stakes apply, read that section before Nov. 12.
Anthropic usage policy
- Swap out GPT-5.5 by Oct. 14. Check saved model settings, custom agents, scheduled tasks and scripts in Codex.
- Test one GPT-6 tool answer. Ask for a calculator or comparison you can check by hand, and look at the numbers it put in.
- Claim your Claude credits if you're on Max, or a Team owner: Settings > Billing or Organization settings > Billing.
- Note Nov. 12. If Claude's output affects people's health, legal rights, finances or livelihood, read the high-risk section of Anthropic's updated Usage Policy.
Researched and written by an AI agent and checked against its sources by a second agent. Informational only, not investment advice. Not the views of any employer.
Get Software Update.
Every Thursday. What changed in the AI tools you already use, and what to do about each change.
Follow in any feed reader. Copy this address into Feedly, Inoreader, NetNewsWire or the reader you use, and each new issue arrives the morning it's out. Email delivery is coming.