Most comparisons quote US prices and test the free apps. This one covers the two team plans an Australian business would actually buy: AUD prices and GST, where your data goes, what the Privacy Act asks of you, what independent tests show and a one-week trial you can run on your own work.
Short answer
Neither tool wins outright, and the independent tests split. ChatGPT Business is the stronger pick if you need image generation, which Claude doesn’t offer, or research reports, which one hands-on reviewer found more useful than Claude’s. Claude Team is the stronger pick if you work through long contracts and big PDFs, where Anthropic documents a larger context window (though one of its own pages gives a smaller figure), or do finance work, where Claude’s models scored above OpenAI’s on Vals AI’s agentic benchmarks. Both connect to Google Workspace, Microsoft 365 and Slack, both include a coding tool and an agent for longer tasks, and neither trains on business plan data by default. ChatGPT Business Standard is AU$28 per user a month billed annually, with no GST wording, and Claude Team Standard is AU$34 including 10% GST, or AU$30.91 before GST, so ChatGPT is cheaper on list price by roughly 9% to 18%, depending on whether OpenAI’s price includes GST. Claude Team stores data in the US. ChatGPT Business now lets you pick a storage region at checkout, but OpenAI doesn’t list which regions Business gets, and it still keeps a copy of prompts and responses in the US for abuse monitoring. Pick on your tools, your data rules and a one-week trial on five real tasks.
A note on fairness. AppliedAgents builds with both tools, isn’t a reseller or partner of either company and earns nothing when you subscribe. An Anthropic model helped draft this page, so every claim comes from the vendor’s own page or a named independent source, checked on 6 November 2026 from an Australian connection, with prices rechecked on 10 November 2026.
What you are actually choosing between
For a business, the real choice is between Claude Team and ChatGPT Business, the two self-serve team plans.
- Claude Team is Anthropic’s plan for 2 to 150 people. You can mix Standard and Premium seats, and every seat includes Claude Code, Anthropic’s coding tool, and Claude Cowork, its agent for multi-step tasks.
- ChatGPT Business is OpenAI’s plan for 2 to 200 people. Every Standard and Premium seat includes ChatGPT, ChatGPT Work (“a separate ChatGPT experience for longer research and deliverable tasks”) and Codex, its coding tool.
- Personal plans such as Claude Pro (AU$34 a month) and ChatGPT Plus (AU$30 a month) are for one person, with different training settings (see the data table) and no shared billing or admin controls.
- Enterprise on both sides adds SCIM, audit logs and retention settings. Claude Enterprise is shown to Australians as US$20 per seat a month billed annually, 20-seat minimum, plus usage at API rates. ChatGPT Enterprise is “Custom pricing”.
Claude Team and ChatGPT Business prices in Australia
ChatGPT Business is cheaper on list price, at AU$28 per Standard seat a month billed annually against Claude Team’s AU$34, though the size of the gap depends on whether OpenAI’s price includes GST. Claude’s Australian pricing page says “Prices include 10% GST”. OpenAI’s shows “A$” prices and says nothing about GST or tax on either tab.
| Seat and billing period | ChatGPT Business (GST status not stated) | Claude Team (includes 10% GST) | Claude Team before GST (our arithmetic) |
|---|---|---|---|
| Standard, billed annually | AU$28 | AU$34 | AU$30.91 |
| Standard, billed monthly | AU$35 | AU$43 | AU$39.09 |
| Premium, billed annually | AU$140 | AU$170 | AU$154.55 |
| Premium, billed monthly | AU$175 | AU$213 | AU$193.64 |
| Minimum team, two Standard seats, billed annually | AU$56 | AU$68 | AU$61.82 |
The before-GST column comes from Anthropic’s own pricing endpoint. Annual Claude seats are billed up front: AU$370.91 before GST per Standard seat for the year.
The GST comparison problem
ChatGPT Business comes out cheaper whichever way the GST question falls, by our arithmetic. If OpenAI’s AU$28 excludes GST, it compares with Claude’s AU$30.91 before GST (or AU$30.80 against AU$34 once GST is added to both), so ChatGPT is about 9% cheaper. If AU$28 already includes GST, ChatGPT is about 18% cheaper. Both companies list identical US prices (US$20 and US$25 for Standard, US$100 and US$125 for Premium), so the gap comes from how each converts to AUD and shows tax. What the GST question does change is your invoice, so check the total at the ChatGPT checkout before you enter a card.
If your business is registered for GST, the ATO says to give an overseas supplier your ABN and state that you are registered. GST then doesn’t apply to the imported service, and your business, not the supplier, works out whether any GST is payable under the reverse charge. We haven’t confirmed that either checkout removes GST for an ABN: the OpenAI tax ID page we read covers VAT, and we didn’t read Anthropic’s. Check your first invoice and ask your accountant. This is information, not tax advice.
Seat rules on both plans
- Minimums and caps. Anthropic says “Team plans require a minimum of two members”, up to 150. OpenAI says “A workspace requires at least two paid seats”, up to 200 for workspaces created from 24 August 2026, though “Some customers can start a new ChatGPT Business workspace with 1 paid Standard or Premium seat” and this is “not broadly available”.
- Adding and removing seats. OpenAI charges added seats immediately, prorated, applies reductions from the next billing cycle and says “Unused ChatGPT seats are generally non-refundable”. Anthropic lets you add seats mid-term, with reductions “at your next renewal” and “no prorated refund”.
- API credits. Both include pooled credits for using the models in your own software. Most small businesses won’t use them, and we haven’t verified whether Claude’s pool is in AUD for Australian workspaces.
What it costs for two people and for five
| Team | ChatGPT Business | Claude Team |
|---|---|---|
| Two people, billed annually | AU$56 | AU$68 (AU$61.82 before GST) |
| Two people, billed monthly | AU$70 | AU$86 (AU$78.18 before GST) |
| Five people, billed annually | AU$140 | AU$170 (AU$154.55 before GST) |
| Five people, billed monthly | AU$175 | AU$215 (AU$195.45 before GST) |
Annual billing saves about 20% on both, but locks you in for a year with no prorated refund on either side, so trial on monthly billing first. Our guide to AI implementation costs in Australia covers what sits around the subscription, such as setup and training time.
Where your data goes
Neither team plan trains on your data by default, Claude Team stores it in the US, and ChatGPT Business now lets you choose a storage region at checkout.
| Question | Claude Team | ChatGPT Business |
|---|---|---|
| Is team data used to train the models? | No by default. “By default, we will not use your inputs or outputs from our commercial products (e.g. Claude for Work…) to train our models.” The exception is thumbs up or down feedback, kept “for up to 5 years”. Owners can switch feedback off. | No by default. “OpenAI won’t train on your workspace’s data.” Response feedback “does not permit training on the associated conversation” on Business. |
| What about the personal plans? | The pricing table lists “Model training: Opt-out” for Free, Pro, Max and Ultra. The toggle is under Settings, Privacy. The help pages we read don’t state the default in words. | “On personal workspaces with ChatGPT Free, Plus, or Pro, data sharing is enabled by default.” |
| Where is it stored? | In the United States. Anthropic’s page for commercial products says “Note that data is stored in the US.” No storage region choice on Team. Anthropic adds that it may process data for internal purposes, such as safety-related review, in countries where it or its affiliates operate. | Starting to vary. OpenAI says “ChatGPT Business lets you choose where your workspace’s customer content is stored”, but only “Where available” and “during checkout for a new Business workspace”: “You cannot turn on data residency or change the region for an existing Business workspace” (checked 18 November 2026). It doesn’t list which regions Business gets. The choice covers storage only: it “does not include inference residency or determine where ChatGPT processes requests”. We haven’t verified that Australia is one, and OpenAI’s pricing table still shows data residency as Enterprise only. If you choose a region outside the US, “a copy of every prompt and response is also stored in the United States for a limited time for safety, abuse monitoring, and enforcement.” |
| Can you choose Australia for the AI processing? | No. By default Anthropic “may route customer traffic to select countries in the US, Europe, Asia and Australia”, and you can’t pick a location on Team. “US-only inference” exists only on usage-based Enterprise, at 1.1x usage rates. | No. Choosing a storage region doesn’t cover processing. Inference residency is Enterprise and Edu only, in Europe, the United States and the United Arab Emirates. |
| Is there an Australian route at all? | Anthropic’s regional compliance page lists Australia, but through AWS Bedrock and GCP Vertex, which are developer platforms, not the Claude Team app. | Yes for new ChatGPT Enterprise and Edu workspaces, which can store content at rest in Australia, one of ten listed regions. |
| How long is it kept? | Until deleted. Deleted chats are “Deleted from our back-end storage systems within 30 days”. Content flagged for Usage Policy violations can be kept “up to 2 years”. Custom retention settings are Enterprise only. | OpenAI’s enterprise privacy page contradicts itself. Its Business FAQ says “Your workspace admins can control how long your data is retained”, while its summary lists retention control only for “ChatGPT Enterprise, ChatGPT for Healthcare, ChatGPT Edu”. Deleted conversations are removed within 30 days. |
| Can the workspace admin read staff chats? | We didn’t check this for Claude Team. | Yes. Business admins “can view, access, export, and delete end user conversations in the workspace”. Usage analytics are separate and don’t show transcripts. |
| Connected tools | The training exclusion covers “raw content from connectors (e.g. Google Drive)”. Connected third-party services may sit outside the US. | Residency doesn’t cover “Apps & MCP, Web Search”. Third parties follow their own terms. |
The practical reading: on both team plans your prompts, uploads and answers are handled offshore, and neither lets you choose Australian processing. If some of your information must stay in Australia, neither plan is the answer for that information. Our guide on what customer information to keep out of AI tools helps you sort what can go in.
The Privacy Act and APP 8 in plain English
If the Privacy Act covers your business, APP 8 makes you accountable for what an overseas AI provider does with the personal information you give it. This is general information, not legal advice. The Office of the Australian Information Commissioner (OAIC) says “Most small businesses are not covered by the Privacy Act 1988, but some are.” The usual line is annual turnover of A$3 million or less, with exceptions that include health service providers, credit reporting bodies and businesses with anti-money laundering reporting obligations. The OAIC page we read was last updated in October 2024 and we haven’t checked whether reforms have changed the exemption since, so first check whether the Act covers you.
If it does, APP 8.1 says that “before an APP entity discloses personal information about an individual to an overseas recipient, the entity must take such steps as are reasonable in the circumstances to ensure that the recipient does not breach the APPs”. Under section 16C, your business “is accountable for an act or practice of the overseas recipient”. In plain terms: if you paste a customer’s details into a tool that stores them in another country, you stay responsible for what happens to them.
The OAIC guidelines say handing data to an overseas cloud provider “for the limited purpose of performing the services of storing” it may count as a use rather than a disclosure, but only where a binding contract limits what the provider can do and leaves you in effective control. An AI service does more than store data: it processes it to produce answers. We haven’t assessed whether either vendor’s terms meet that test, so don’t assume the carve-out applies. The OAIC’s AI guidance recommends “that organisations do not enter personal information, and particularly sensitive information, into publicly available generative AI tools”. Our one-page AI policy template gives you a written rule for what stays out, which is the most useful control on either plan.
Admin and security controls compared
Both team plans give you single sign-on, and both keep SCIM, audit logs and full role-based access controls for Enterprise.
| Control | Claude Team | ChatGPT Business |
|---|---|---|
| Single sign-on | Yes, with domain verification and just-in-time provisioning | Yes, SAML SSO, with MFA and domain verification |
| SCIM user provisioning | Enterprise only | Enterprise only |
| Audit logs or compliance API | Enterprise only. The pricing table says “Audit logs: No” for Team | Enterprise only (Compliance API and Logs Platform) |
| Role-based access | Anthropic’s pages disagree. The pricing table says Team “Role-based access: No”, while the Team help article lists “Role-based permissioning” | Business has admin roles (owners, admins, analytics viewers, members). “Role-based access controls” are listed for Enterprise only |
| Retention controls | Enterprise only | Contradictory, see the data table above |
| Certifications (sources differ, see below) | Anthropic’s Trust Center lists SOC 2 Type 2, ISO 27001, ISO 42001, CSA STAR and NIST 800-171 for Claude Team. HIPAA is Enterprise only | OpenAI’s pricing table ticks SOC 2 Type 2 for Business, and its “ISO 27001, 27017, 27018, and 27701 certified” line only for Enterprise |
| Custom connectors (MCP) | Only Owners and Primary Owners can add them | Only admins and owners can create and publish them |
The certification row isn’t like for like. Anthropic’s list comes from its Trust Center, while OpenAI’s comes from its pricing page, which points to a separate trust portal we didn’t read. Each portal may list more than the summary we quote, so if a certification matters to a client contract, ask both vendors for the report. For a business of two to twenty people, SSO plus a written policy usually covers it. If an insurer, client or regulator asks for audit logs, you are looking at Enterprise on either side.
What each does better day to day
ChatGPT Business makes images and Claude Team doesn’t, each has one documented edge for heavier work, and most everyday jobs are close.
| Job | Better fit | Why |
|---|---|---|
| Making images for ads, social posts or mock-ups | ChatGPT | Image generation is included on Business. “Claude doesn’t generate photos or illustrations”; it makes HTML and SVG charts and diagrams instead. |
| Reading a long contract, tender or report in one go | Lean Claude, on vendor figures | Anthropic’s per-model page lists up to 1 million tokens of context on its current models, but its Team plan page still says “200k context window”. ChatGPT Business lists 54K for Instant (about 40 pages) and 256K for Reasoning (about 320 pages). If the 200K figure applied, ChatGPT’s Reasoning mode would hold more. |
| Research reports you can hand to someone | Lean ChatGPT, on one review | Both have a research mode that uses the web and your connected apps. ChatGPT’s deep research reports download as Markdown, Word or PDF. One hands-on reviewer, using older models, found ChatGPT’s reports “usually” more useful. |
| Handing off a multi-step task | Either | Claude Cowork runs on the desktop on all paid plans, with web and mobile “in beta on Team”, and can run scheduled tasks. ChatGPT Work is on every Business seat, plus workspace agents you can share and schedule. |
| Working inside Google Workspace or Microsoft 365 | Either | Both connect to Google Workspace, Microsoft 365 and Slack. Claude can edit Google Docs, Sheets and Slides live and works inside Excel, PowerPoint, Word and Outlook. ChatGPT has Excel, Word, PowerPoint and Google Sheets extensions. Both now take meeting notes, in beta and in different ways: Claude Meetings joins Google Meet calls as a visible participant on Team and Enterprise plans, and ChatGPT’s Meetings plugin takes notes from the Mac desktop app on Pro and Business plans without a bot joining. Get everyone’s consent first; our AI note taker consent guide covers what to say. |
| Someone on the team writes code | Either | Claude Code is on every Team seat. Codex is on every Business seat, with Codex Cloud available. |
| Microsoft Copilot is already in the mix | See our Copilot guide | Our ChatGPT Business vs Copilot guide covers that choice. |
Files, agents and limits the table leaves out
Files. Claude chat takes files up to 500MB, 20 per chat, and PDFs up to 1,000 pages, with visuals analysed for the first 100 pages only. ChatGPT Business takes 512MB per document, 2 million tokens per text file and 80 files every three hours, with 25GB of storage per user. Both accept big files. What differs is how much the model holds in view at once, and since Anthropic’s own pages disagree on that, test your longest real document on both.
Agents. Both agents have a built-in browser: it is on by default in Cowork on Team, and ticked for ChatGPT Business. Computer use, where the agent operates apps on your machine, is “in beta for Pro, Max, and Ultra plans” on Claude, and Team isn’t listed. We haven’t verified what desktop computer use ChatGPT Business offers. OpenAI retired the older ChatGPT agent and now says: “Use ChatGPT Work for longer, multi-step tasks and finished deliverables.”
Usage limits. Both meter usage and change the rules often. A Claude Team Standard seat has “1.25x the Pro plan’s per-session usage allowance” and a Premium seat “6.25x”, with five-hour and weekly limits and extra usage credits you can buy. On ChatGPT Business, Premium seats “include 5x more usage than Standard seats and are not subject to a five-hour usage limit”. From December 2026, OpenAI says Business usage “will be shared across ChatGPT and Codex”, measured in tokens, and “Instant chat will not draw from your usage”. If someone keeps hitting limits, a Premium seat for that person is the first step on either side.
What the independent tests say
The independent tests don’t name a single winner: two benchmarks put Claude’s models ahead of OpenAI’s, and the one hands-on review preferred ChatGPT.
- PCMag Australia, hands-on review (updated 24 August 2026). One reviewer used the consumer apps for research, images, files, writing, reasoning, coding and integrations, comparing GPT-5.6 with Claude Opus 5, both since replaced. The reviewer preferred ChatGPT: Claude’s research reports “usually aren’t quite as useful”, ChatGPT understood complex images better, and GPT-5.6 was the pick for casual coding. The review noted Claude’s larger context window but said ChatGPT’s compaction makes the difference unnoticeable in most situations, and on writing said “There’s Nothing Either Can’t Write”. Strength: low as a ranking. It is one tester on older models and consumer plans, the site carries deals and affiliate links, and PCMag discloses that its parent company is suing OpenAI over copyright.
- LMArena Text Arena (27 October 2026, about 9 million votes, 418 models). People put the same prompt to two unnamed models and vote for the better answer. Google’s Gemini 4 models are first and second. Claude Opus 5.5 is third at 1509 and Claude Fable 5.1 fifth at 1504; GPT-6.1 Sol is 27th at 1481 and GPT-6 Astra 37th at 1476, each give or take about 6 to 7 points. Strength: medium for general preference, weak for business tasks, because voters type whatever they like and writing style sways votes.
- Vals AI, Vals Index and Finance Index (5 November 2026). Private tests of agentic finance, coding, legal and tax tasks, run on the models through the API. Overall: Claude Sonnet 5.5 70.85%, Gemini 4 Carbon 70.45%, Claude Opus 5.5 70.18%, GPT-6 Astra 65.02%, GPT-6.1 Sol 63.94%. On the Finance Index, built on US SEC filings, Gemini 4 Carbon leads at 67.05%, Claude Opus 5.5 scores 65.59% and both GPT-6 models about 56%. Vals notes the Claude scores “include tasks served by a fallback model after a refusal”. Strength: medium to high for agentic US finance and legal work, weaker for Australian small business tasks or the apps. We didn’t audit Vals’ funding or independence.
The evidence is lopsided, and we haven’t padded it to look even. Two of the three sources put Claude’s models ahead of OpenAI’s, but both test models rather than the apps you would buy. The only source that used the apps preferred ChatGPT, and it is the weakest method of the three. OpenAI’s GDPval and Anthropic’s Australian usage data are vendor-run, so we left them out. All of these models will be replaced within months, so the test that counts is the one you run on your own work.
Which should you choose?
Start from the tools you already use and the jobs you do most, then let a trial decide.
| Your situation | Start with |
|---|---|
| You run on Google Workspace or Microsoft 365 and mostly need writing, summaries and email help | Either. Both connect to both suites. Trial both on monthly billing and keep the one your team opens without being told. |
| Marketing needs generated images | ChatGPT Business. Claude doesn’t generate images. |
| You want research reports to hand on to clients or staff | Lean ChatGPT Business on one reviewer’s experience, but run the same question through both research modes. |
| Your work lives in long contracts, tenders and big PDFs | Lean Claude Team on its documented context window, but test ChatGPT’s Reasoning mode on the same file. |
| Finance or legal analysis | Lean Claude Team on Vals AI’s benchmarks, which test API models on US material. Check APP 8 and your professional obligations before client data goes in. |
| Some information must stay in Australia | Neither team plan, for that information. ChatGPT Enterprise can store content in Australia for new workspaces, with processing still offshore. Talk to your adviser first. |
| It is just you, or you and one other person | Both require two seats, though OpenAI offers one seat to some new workspaces. A personal plan can work for one person if you switch training off and keep customer details out. |
Try both for a week
Two Standard seats on each, billed monthly, for one week of real work will tell you more than any review: AU$70 for ChatGPT Business and AU$86 including GST for Claude Team, by our arithmetic. Neither gives a prorated refund if you cancel seats early, so set a reminder for the day before each renewal.
- Pick the two people who will use it most, not the two who are most curious.
- Buy two Standard seats on each, billed monthly. Connect the same tools to both, such as Google Drive or Microsoft 365.
- Run the five tasks below on both, with the same brief each time. Our prompting guide has a template.
- Score each task the same day, while you remember how much you had to fix.
- On day seven, add up the scores and ask which app each person opened on their own.
- Cancel the loser before renewal, then write the rule for what the winner may be given, using our AI policy template.
Five tasks that reflect real business work
- The email thread. Paste a customer thread of 15 to 20 messages and ask for a summary of what was agreed and a draft reply in your tone.
- The long document. Upload a 40-plus page PDF you know well, such as a supplier contract, tender or award. Ask for a one-page brief with page references, then check three of them.
- The proposal. Give it your rough notes and one past proposal, and ask for a new quote cover letter. Count how many sentences you change.
- The spreadsheet. Upload a sales or jobs export and ask for the three things that changed this quarter, with a chart. Check the numbers against the sheet.
- The research question. Ask each research mode the same question, such as a comparison of three suppliers, with sources. Open two of the sources.
Scoring sheet (copy one per task)
Task: [e.g. one-page brief of the supplier contract]
Who ran it: [name] on [date]
• Time it took before AI: [minutes]
• ChatGPT Business: [minutes] | Accuracy 1 to 5: [ ] | Sounded like us 1 to 5: [ ] | Fixes needed: [ ]
• Claude Team: [minutes] | Accuracy 1 to 5: [ ] | Sounded like us 1 to 5: [ ] | Fixes needed: [ ]
• Did either make something up or miss something important? [which, what]
• Did we paste in anything that should have stayed out? [yes/no, what]
• Who would use this every week? [names]
• Worth paying for? [yes/no, why]
If one tool saves a person an hour a week on real work, it has paid for its seat. If both do and the scores are close, pick the one your team opened unprompted. If neither does, look at the tasks and briefs before you decide AI isn’t for you. If you want help, our AI implementation service covers choosing, rollout, connectors and training.
Common questions
Short answers to the questions people most often ask about Claude and ChatGPT for business.
Is ChatGPT or Claude better for my business?
Neither is better across the board. Decide on the tools you already use, whether you need image generation (ChatGPT only), your data rules and a one-week trial on real tasks. The independent tests split: Claude’s models score higher on agentic finance and legal benchmarks, a hands-on reviewer preferred ChatGPT for research and images, and crowd preference puts Google first.
Is Claude the best AI for businesses?
No single tool is “best”. Claude’s current models lead the Vals Index (5 November 2026), but Google’s Gemini 4 leads other tests and one hands-on review preferred ChatGPT. Claude Team has clear limits too: no image generation, data stored in the United States, and audit logs only on Enterprise.
Which ChatGPT is best for business?
For a team, ChatGPT Business: AU$28 per user a month billed annually (checked 10 November 2026, GST status not stated), usually a two-seat minimum, and no training on your workspace data by default. Plus is a personal plan with data sharing on by default. Enterprise adds Australian data storage for new workspaces, SCIM, role-based access controls and the Compliance API, at custom pricing.
Is Claude better than ChatGPT for finance professionals?
On Vals AI’s Finance Index (5 November 2026), Claude Opus 5.5 scored 65.59% against 55.81% for GPT-6.1 Sol, but Gemini 4 Carbon was top at 67.05%. The tasks use US SEC filings and API models, not the apps, and the Claude scores include tasks served by a fallback model after a refusal. It doesn’t show that Claude is better for an Australian bookkeeper or accountant. Neither tool should handle client financial data until you have checked APP 8 and your licence obligations.
How much does Claude cost in Australia?
As shown to Australian visitors on 10 November 2026, including 10% GST: Claude Pro is AU$34 a month, or AU$340 a year billed up front. Team Standard is AU$34 per seat a month billed annually or AU$43 monthly. Team Premium is AU$170 per seat a month billed annually or AU$213 monthly. Max is AU$170.01 a month and Ultra AU$340 a month, monthly billing only.
Do these prices include GST?
Claude’s Australian pricing page says “Prices include 10% GST”. OpenAI’s pricing page doesn’t say either way, so check the total at checkout. If you are registered for GST, the ATO says to give an overseas supplier your ABN and state that you are registered, after which the reverse charge rules apply. Ask your accountant how to treat the invoices.
Can I keep my data in Australia?
Not reliably on these two team plans. ChatGPT Enterprise can store content at rest in Australia for new workspaces. ChatGPT Business lets you pick a storage region at checkout, but OpenAI doesn’t list the regions and keeps a US copy for abuse monitoring. Claude Team stores data in the United States. Neither team plan lets you choose Australian processing.
Sources
Every claim above comes from these pages, checked on 6 November 2026.
Anthropic:
- Checked 16 November 2026: Introducing Claude Meetings (12 November 2026), More ways to work with Claude (10 November 2026) and Claude for Google Workspace
- Claude pricing (Australian render, “Prices include 10% GST”) and supported regions
- Claude Team plan (seat minimum, usage multipliers, US prices), purchase and manage seats on Team plans and context windows by model
- Training on commercial product data, retention for commercial products, where data is stored and processed, consumer data retention and US-only inference
- Anthropic Trust Center and regional compliance
- Claude Cowork, computer use, Research, Google Workspace connectors, custom connectors (MCP), file limits and Can Claude produce images?
OpenAI:
- Checked 16 November 2026: The Meetings plugin in ChatGPT and ChatGPT Record
- ChatGPT pricing (Australian render), ChatGPT Business FAQ, single-seat Business workspaces, managing billing and seats and supported billing currencies
- Checked 17 November 2026: where your ChatGPT Business content is stored and data residency and inference residency
- Enterprise privacy, data controls FAQ and Business data and training
- ChatGPT Business models and limits, understanding ChatGPT and Codex usage limits, Codex with ChatGPT plans, custom MCP connections, file upload limits and ChatGPT agent (retired)
Australian regulators:
- OAIC, APP 8 guidelines, Chapter 8 (updated 3 October 2025), guidance on privacy and commercially available AI products and privacy checklist for small business (updated 21 October 2024)
- ATO, Australian business importing goods and services (updated 11 September 2025)
Independent tests:
- PCMag Australia, ChatGPT vs. Claude: After Testing Both, I Have a Clear Favorite (updated 24 August 2026)
- LMArena Text Arena leaderboard (27 October 2026)
- Vals AI, Vals Index and Finance Index (5 November 2026)
Prices were rechecked on 10 November 2026; plan rules, data policies and feature lists were checked on 6 November 2026 (ChatGPT Business storage on 17 November 2026), all from an Australian connection. Claude’s prices include 10% GST; OpenAI’s pricing page does not state GST treatment. Several features on both sides are in beta or rolling out, and OpenAI’s usage limits change in December 2026.
Practical thinking on AI, automation and the work behind a business.



