How much do Claude Pro and ChatGPT Plus cost?
The same, give or take an annual discount. All prices below were taken from the vendors' own pages on October 22, 2026, in US dollars, before tax.
| Plan | Price | Billing | Source |
|---|---|---|---|
| Claude Free | $0 | — | Claude pricing |
| Claude Pro | $20 a month, or $17 a month billed annually ($200 up front) | Monthly or annual | Claude pricing |
| Claude Max | From $100 a month, at 5x or 20x Pro's usage | Monthly | Claude pricing |
| ChatGPT Plus | $20 a month | Monthly | OpenAI Help Center |
| ChatGPT Pro | $100, $200 or $500 a month | Monthly | OpenAI Help Center |
Neither subscription includes API usage. If you call the models from your own code, you pay per token on top: $4 and $20 per million input and output tokens for Claude Opus 5.5, and $10 and $50 for GPT-6 Astra.
Which plan gets you the models that built our projects?
This is the question that matters if you liked a build on this site. Each row shows where the model can be used on a consumer plan, according to the vendor's own pages.
| Model | Cheapest plan that includes it | Notes |
|---|---|---|
| Claude Opus 5.5 | Claude Pro | Not on Free |
| Claude Fable 5.1 | Claude Max for regular use | Pro only through extra usage credits; Max up to half of weekly limits |
| GPT-6 Astra | ChatGPT Plus | OpenAI rolled it out to Plus, Pro, Business and Enterprise |
| GPT-6.1 Sol | ChatGPT Pro and Business list it | OpenAI's Plus page says to check the model picker |
The Fable row is the surprise. Claude's pricing table lists Fable on Pro only through usage credits, and on Max 5x and Max 20x for up to 50% of weekly limits. If Fable 5.1's restrained, editorial builds are what you want, Max is the realistic plan.
What do you get on the free plans?
Enough to try both, not enough to build with an agent. Claude's Free plan includes chat on web, desktop and mobile, web search, file creation and Artifacts you can preview, but neither Claude Code nor Claude Opus 5.5, according to its pricing table. On the ChatGPT side, GPT-6 Instant reached free users on October 8, 2026, a day after paid users, according to Wikipedia's GPT-6 entry, while GPT-6 Astra is for paying plans. Either free plan will write you a single page you can paste into a file. Neither will run a build loop for you.
Which coding agent does each plan include?
Claude Code comes with every paid Claude plan: Claude's pricing table marks it "No" on Free and "Yes" on Pro, Max 5x and Max 20x. OpenAI's Help Center lists Codex among the features of ChatGPT Pro, and OpenAI's business pricing page includes Codex with Business seats, which cost $20 a month billed annually or $25 monthly per user. For Plus, OpenAI describes something narrower: Plus subscribers can let participating tools draw on their included allowance through Sign in with ChatGPT.
In practice, then, $20 buys you a full coding agent on the Claude side without question. On the OpenAI side, check what your plan shows before you count on Codex at that price.
How fast will you hit the usage limits?
Faster than with chat, if you use an agent. Every agent step re-reads your project, so a single build can process millions of tokens. Our Claude Opus 5.5 landing page read about 10 million input tokens, and two maximum-effort Claude Code sessions running side by side exhausted a five-hour usage window in about two hours during our runs. Anthropic raised five-hour limits on Pro, Max, Team and Enterprise plans with the Opus 5.5 launch, according to its announcement, but agent work still drains them quickly.
The GPT agents were far more frugal on our briefs: their builds used 0.6 to 2.7 million input tokens and finished in 7 to 25 minutes. If you mostly want quick, complete builds and will iterate, that efficiency stretches a subscription further.
What would the same builds cost pay-as-you-go?
Here is the API-equivalent cost of one build per model, from our run logs, at each vendor's list prices. It is a useful yardstick for whether a subscription pays for itself.
| Model | Finished | Wall time | Output tokens | Files | Source size | Runtime errors | API-equivalent cost | Attempts |
|---|---|---|---|---|---|---|---|---|
| Claude Opus 5.5 | Yes | 1h 21m | 285k | 1 | 115 KB | 0 | $13.63 | 2 |
| Claude Fable 5.1 | Yes | 38m 43s | 142k | 1 | 77 KB | 0 | $12.80 | 1 |
| GPT-6 Astra | Yes | 10m 27s | 35k | 1 | 68 KB | 0 | $2.79 | 1 |
| GPT-6.1 Sol | Yes | 7m 26s | 36k | 1 | 67 KB | 0 | $0.54 | 1 |
A Claude build of our dashboard brief costs $12 to $14 at API prices; a GPT build costs $0.54 to $2.79. Two or three Claude builds a week already exceed the price of Claude Pro, which is why subscriptions with generous agent limits are attractive for builders. For the GPT models, pay-as-you-go stays cheap unless you build constantly.
Claude Max or ChatGPT Pro for heavy use?
Claude Max if your heavy use is building, ChatGPT Pro if it is everything else plus some building. Claude Max starts at $100 a month with five or twenty times Pro's usage, regular access to Fable 5.1 and $100 to $200 of monthly API credits included. ChatGPT Pro starts at $100 with Codex and deep research, and its $500 tier adds Ultrafast versions of GPT-6 Astra and GPT-6.1 Sol. Both vendors change usage allowances often: OpenAI's Help Center notes that new Pro 200 subscriptions now come with a lower allowance than before at the same price.
What about team plans?
Both vendors price teams per seat, and both now separate heavy users from light ones. OpenAI's business pricing page lists a standard ChatGPT Business seat at $20 a month billed annually or $25 monthly, and a premium seat at $100 a month billed annually or $125 monthly with five times the usage and no five-hour limit; both include Codex. Anthropic's pricing page describes its Enterprise plan as $20 per seat per month plus usage billed at API rates, so cost scales with the models and tasks a team runs.
For a small team that builds with agents, the difference in billing models matters more than the headline seat price. Per-seat allowances suit steady, predictable use. Usage-based billing suits a team where two people run agents all day and everyone else mostly chats.
Which should you choose?
Choose by the work. If you will hand projects to an agent and want the most thorough first draft, Claude Pro is the better $20, and Max the better $100. If you want a broad assistant and quick, inexpensive builds, ChatGPT Plus is the better $20 and the GPT-6.1 Sol API a cheap way to build more. Before paying for either, watch the two vendors' builds of the brief closest to your work, ideally with the names hidden, and remember that both vendors revise plans and allowances often, so check the date on any comparison, including this one.
Pick the landing page you would pay for, then see which subscription built it.
Take the blind test
Frequently asked questions
Is Claude Pro or ChatGPT Plus better for coding?
For building whole projects, Claude Pro, because it includes Claude Code and Claude Opus 5.5, the model behind the most elaborate builds in our tests. ChatGPT Plus is the better everyday assistant at the same price and includes GPT-6 Astra, but OpenAI lists its Codex agent explicitly only on the Pro and Business plans.
How much do Claude Pro and ChatGPT Plus cost?
As of October 22, 2026: Claude Pro is $20 a month billed monthly, or $17 a month billed annually as $200 up front. ChatGPT Plus is $20 a month billed monthly. Prices exclude tax.
Does Claude Pro include Claude Fable 5.1?
Only through extra usage credits. Claude's pricing page lists Fable as available on Pro via usage credits, and on the Max plans for up to half of the weekly limits. Claude Opus 5.5 is included in Pro.
What is the difference between ChatGPT Pro and Claude Max?
Both are the heavy-use tiers. ChatGPT Pro comes in three monthly plans at $100, $200 and $500, with the $500 plan adding Ultrafast GPT-6 models. Claude Max starts at $100 a month with five or twenty times Pro's usage and includes $100 to $200 a month in API credits.
Which plan did GlanceHall use for its builds?
Subscriptions on both sides, with each model run in its vendor's own agent: Claude Code for the Claude models and Codex for the GPT models. We publish what each build would have cost at list API prices so the comparison does not depend on any plan's limits.