Should I pay for Claude? Making sense of AI tiers

Most surgeons who ask me about AI assume the expensive tier is the right starting point. It usually isn’t. I made the same mistake: I paid for the top option before I understood what I actually needed. This post is the answer I wish I’d had before I spent any money.


The three tiers, in plain terms

Free. Claude’s free tier gives you access to the web interface, iOS and Android apps, and the desktop app. You can chat, search the web, generate text, write and edit documents, run basic code, and use memory across conversations. Usage is limited — you’ll hit a cap after a certain volume of messages, and it resets on a rolling basis. For someone dipping in occasionally, or testing whether Claude is useful at all, Free is a reasonable place to start.

Pro. At around £19 per month (or roughly £17 on annual billing — Anthropic’s prices are set in USD and UK VAT is added at checkout), Pro removes most of the friction. You get significantly more usage, access to all Claude models including the most capable ones, unlimited Projects (the feature that lets you build a persistent workspace with context Claude remembers), and Claude’s Research mode, which does multi-step searches and synthesises sources rather than just answering a single query. Pro also includes Claude Code and Claude Cowork — the desktop agentic tool. For anyone using Claude regularly in clinical or administrative work, Pro is the practical tier.

Max. From around £96 per month (5x Pro usage) or approximately £192 per month (20x Pro usage), Max exists for people who routinely hit Pro limits — developers running code in loops, researchers processing large document sets, teams with high-volume workflows. It also includes early access to new features and priority availability at peak times. For most surgeons, this is not the right tier. If you’re not regularly hitting Pro’s limits, you’re paying for headroom you won’t use.


How does this compare to ChatGPT?

The comparison most people reach for first. ChatGPT Plus sits at the same price point as Claude Pro — around £19 per month including VAT — and includes GPT-5.5 at high usage limits, Deep Research (capped at ten runs per month), and access to tools like Sora and Codex. OpenAI also now offers a mid-tier option at around £96 per month for heavier users, with a top-tier Pro plan at roughly £192 per month.

The pricing structures are broadly similar. The models are not interchangeable — Claude and ChatGPT reason differently, write differently, and perform differently across different tasks. That comparison is explored in detail here. The short version: for writing, complex document work, and tasks where you want consistent tone and careful reasoning, I prefer Claude. For coding and certain technical tasks, GPT has strengths too. Most surgeons using AI for clinical or administrative work will find Pro-level access in either product sufficient. You do not need both.


The honest decision framework

Start with Free. Use it for a few weeks on real tasks — drafting referral letters, summarising guidelines, preparing for a governance meeting, working through a complex consent conversation. If you hit the usage cap regularly and find yourself frustrated, that is the signal to upgrade. If you’re hitting the cap and the work matters, around £19 a month is not a hard decision.

Upgrade to Pro when you want to use Projects properly. Free gives you limited project access; Pro removes that ceiling. Projects are the feature that shift Claude from a chatbot you ask occasional questions to a tool that holds context about your clinical environment, your patients, your writing style, and your working week. That is the capability that changes how useful it is. Setting one up takes about fifteen minutes — here’s how.

Skip Max unless you already know you need it. If you reach Pro and find you’re consistently hitting limits mid-task, reassess. Until then, Max is oversized for the job.


One more thing worth saying: both Claude and ChatGPT change their pricing and feature sets more often than most software products. What I’ve described here is current as of May 2026 — check the Anthropic pricing page before subscribing, because the tiers evolve.

The right starting point is Free. The right upgrade trigger is when you notice the limits.


Not sure which tier is right for your setup? Logged-in members can book a session — 30 minutes (£75) to work out exactly what you need and get it configured, or 60 minutes (£150) to build a proper working setup from scratch.


Notes

  1. Claude pricing: Anthropic. Plans & Pricing. claude.com/pricing. Accessed May 2026.
  2. ChatGPT pricing: OpenAI. ChatGPT Plans. chatgpt.com/pricing. Accessed May 2026.

1 thought on “Should I pay for Claude? Making sense of AI tiers”

Leave a comment