Ashley Hirst Writing on community, artificial intelligence and insurance
Artificial intelligence

Anthropic's Max plan has a trust problem

Promise 4x, charge 2x, get ~1.5x — not a great way to get trust.

Ashley Hirst · 31 August 2026 · 2 minute read

Anthropic have just committed an act of brand and customer trust self-harm that I can’t even begin to understand.

If you’re not an AI nerd, a couple of things to know: Fable is Anthropic’s leading model, and it is still undoubtedly ahead of everything in terms of design. OpenAI’s ChatGPT leader, Sol, is the same standard on pretty much everything else, and possibly better on coding. But OpenAI have been absolutely leading on how much capacity they provide. If you buy their top Max 20x plan as a consumer, the amount of use you get seems to be much higher than Anthropic’s Max 20x offer. And Anthropic have even limited Fable to being 50% of your total usage — no such limit with OpenAI.

Now to today’s news.

If you were offered a product that offered two price points — one with 5x use, one with 20x use — you’d think the second one had four times the usage of the first one, right? Especially if it’s double the price — you’d think that makes sense: a discount for volume. That’s exactly how OpenAI’s product works.

But it’s not how Anthropic’s offer works.

Turns out, in the small print, that what they meant was that the amount you can use within five hours is four times higher. But the total amount you can use in your subscription is only around 1.5x higher. So you’d be better off getting two of the smaller subscriptions.

Infographic comparing Anthropic's Max 20x and Max 5x weekly usage limits

The coding / AI community is, in my view correctly, furious about this. We’ve all been carefully rationing our spend and usage. And the product’s excellent, but the Max 50% usage during the week is frustrating. But this — this is just about trust.

I have zero doubt Anthropic’s lawyers will be able to explain why this is OK. But it feels morally awful. And a brand that is generally trusted and genuinely liked by hardcore users is now universally distrusted by that same group.

This is a company aiming to IPO this year. I’ve no idea how they’re going to rescue this, but they need to.

Here’s Anthropic’s own Fable assessing the situation: Anthropic’s Fable assessment.


Portrait of Ashley Hirst

I work in insurance and write about artificial intelligence, risk and community — Jewish and British. This site collects the writing. More about me.