A study Microsoft paid for puts the monthly cost of an everyday Claude Enterprise user at $205, against $30 for Copilot and $81 for ChatGPT Enterprise, Microsoft said on 16 November. Anthropic says its non-coding Enterprise users average under $75 a month.
Spread over the study’s own workload of 100 tasks a month, that is $2.05 a task for Claude, 81 cents for ChatGPT and 30 cents for Copilot, by our arithmetic.
Microsoft made the claim at its Ignite conference. Copilot itself offers Anthropic’s Claude models, as we reported. For more on the company, see our Microsoft stock page.
Business Pill · The Model and the Harness
An evergreen Business Engineer explainer. It teaches the general idea only and does not discuss this study.
The key insight: Microsoft’s GigaOm study puts an everyday Claude Enterprise user at $205 a month against $30 for Copilot, $2.05 versus 30 cents a task by our arithmetic; Anthropic says non-coding users average under $75.
What Did Microsoft Claim?
“Anything you can do in Claude or ChatGPT, you can do better with Copilot, and for less,” Judson Althoff, who runs Microsoft’s commercial business, said at its Ignite conference, The Next Web reported.
He was citing research Microsoft “commissioned and funded” with the analyst firm GigaOm, titled Demystifying the TCO of Enterprise AI. Microsoft ran the tests and the cost analysis; GigaOm helped design the research and validated the method and results, Microsoft said.

What the Study Assumes
Testing ran in October. Each product did the same 37 prompts for the same made-up company. A modeled month is 100 tasks for an everyday user and 140 for an advanced one, including 20 agentic workflows.
Prices are October list prices, or assumed prices where none is published, with no negotiated discounts. Microsoft says rivals’ seat prices may be $10 lower than Copilot’s, but usage charges erased that gap within the first week of modeled use.
Human reviewers graded every answer against a checklist, and Microsoft says the three products delivered “broadly similar quality”. It tells readers to treat the findings “as directional” and says actual pricing may vary.
How the Gap Changes for Heavy Users
For advanced users the study puts Copilot at $96 a month, ChatGPT at $149 and Claude at $299. Per modeled task that is 69 cents, $1.06 and $2.14, by our arithmetic.

So the relative gap narrows as work gets heavier, even though the dollar gap between Claude and Copilot widens from $175 to $203. Claude goes from 6.8 times Copilot’s cost to 3.1 times, and ChatGPT from 2.7 times to 1.6 times, our arithmetic. Copilot’s own bill more than triples, because agentic work runs on usage-based Copilot Credits.
What Anthropic Says
Anthropic told The Next Web that Claude Enterprise users who don’t write code average under $75 a month. That is at least $130, or about 63%, below the study’s $205, by our arithmetic.
The two figures measure different things. Anthropic’s is an average of what real customers spend; Microsoft’s is a modeled workload at list price. GigaOm, which Microsoft paid, validated the study’s method and results, Microsoft said.
The Structural Read
As we read it, the fight is over usage fees, not seat prices: the study’s own words say rivals’ seats can be cheaper until usage charges are added.
As we read it, a vendor-funded study whose post does not list its price inputs reads as a sales argument more than a benchmark, and Anthropic’s lower figure comes without a published method either.
Judson Althoff, Microsoft, at Ignite, via The Next Web, 18 Nov 2026
“Anything you can do in Claude or ChatGPT, you can do better with Copilot, and for less.”
Three Implications
FOR BUYERS The study uses list prices with no discounts; Microsoft calls its findings directional and says actual pricing may vary.
FOR ANTHROPIC It says non-coding Claude Enterprise users average under $75 a month, against the study’s $205.
FOR MICROSOFT Copilot’s modeled cost rises from $30 to $96 a month for advanced users, as agentic work moves onto usage-based Copilot Credits.
What Is Not Established
Microsoft’s post does not give the seat prices, usage rates, plan tiers or models behind each total, so the build-up of each figure cannot be checked from it. The full GigaOm report was not readable to us.
The Bottom Line
Microsoft’s own numbers put an everyday Claude task at $2.05 against 30 cents on Copilot, by our arithmetic, but the study is vendor-funded and Anthropic says real non-coding users spend under $75 a month. What to watch is whether Microsoft publishes the full price build-up, and whether enterprise buyers see the same gap on their invoices.
97,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. The per-user costs, workload and method are from Microsoft’s 16 November Copilot blog post on research it commissioned and funded with GigaOm; we could not read the full GigaOm report. Althoff’s quote and Anthropic’s under-$75 figure are as reported by The Next Web. Costs per task, ratios and the 63% gap are our arithmetic on those figures.
Sources: Microsoft Copilot Blog, ‘What AI at work really costs’, Nicole Herskowitz, 16 Nov 2026 · GigaOm, ‘Demystifying the TCO of Enterprise AI’ (commissioned and funded by Microsoft) · The Next Web, ‘Microsoft says Copilot beats Claude and ChatGPT on price’, 18 Nov 2026 · FourWeekMBA on Copilot offering Claude models, Oct 2026









