AI Pricing Models: How to Pick the Right Plan for Your Business
Short answer Most small businesses should start on a flat monthly subscription: $20 a month buys ChatGPT Plus (with Codex) or Claude Pro (with Claude Code). Pay per token through the API only when software you build runs AI on its own. Move up a tier when you hit usage limits every week, not before.
Want help applying this to your business? Talk it through on a free 30-minute call.
Book a callWhich AI pricing model fits a small business?
A flat monthly subscription, for almost every owner. You pay a set price per person, use the chat app and its built-in agent, and the worst case is hitting a limit, not a surprise bill. The other two models fit narrower jobs.
| Pricing model | You pay | Fits | Risk |
|---|---|---|---|
| Subscription with limits | A fixed price per person per month | People typing into ChatGPT, Claude, Codex or Claude Code | You hit a usage cap and wait or upgrade |
| Subscription plus credits | The plan, plus extra credits when you run out | Busy months on a plan you otherwise like | Top-ups add up if you buy them every month |
| Per token (API) | For every piece of text the AI reads and writes | Software that calls AI on its own: a website chatbot, a nightly report | The bill grows with use, with no natural ceiling |
A token is a small chunk of text, often part of a word. Anthropic's pricing docs put it simply:
"Tokens are pieces of text that models process." (Claude API Docs: Pricing)
API prices are quoted per million tokens, written "MTok". Text the AI writes (output) costs several times more than text you send it (input).
What ChatGPT and Codex cost
Codex, OpenAI's coding and task agent, comes with every ChatGPT plan. OpenAI's Codex pricing page says:
"ChatGPT Work and Codex are included in your ChatGPT Free, Go, Plus, Pro, Business, Edu, or Enterprise plan" (ChatGPT: Codex pricing)
| ChatGPT plan | Price (US dollars, checked November 16, 2026) | Codex use |
|---|---|---|
| Free | $0 | Limited |
| Go | $8 a month | Limited |
| Plus | $20 a month | Capped every five hours; extra credits for sale |
| Pro | $100, $200 or $500 a month | No five-hour limit |
| Business | $25 per user a month, or $20 billed yearly; 2 users minimum | Shared workspace; credits for sale |
| Enterprise, Edu | Contact sales | Flexible pricing |
Source: ChatGPT: Codex pricing. Plus users get an estimated 17 to 180 messages per five hours on GPT-6.2 Sol, fewer on the bigger GPT-6.2 Astra, and far more on smaller models. Weekly limits may also apply.
You can also run Codex with an API key instead of a plan. Then OpenAI says you "Pay for Codex usage based on API pricing". The flagship GPT-6.2 Astra costs $8 per million input tokens and $40 per million output tokens; GPT-6.2 Sol costs $2 and $10; the small GPT-6 Luna costs $0.10 and $0.50 (OpenAI API Docs: Pricing).
What Claude and Claude Code cost
Claude also sells subscriptions first, and Claude Code comes with every paid plan.
| Claude plan | Price (US dollars, checked November 16, 2026) | Usage |
|---|---|---|
| Free | $0 | Limited; no Claude Code |
| Pro | $20 a month, or $200 a year | More than Free |
| Max | $100 a month | 5 times Pro, plus $100 a month in API credits |
| Ultra | $200 a month | 10 times Pro each week, no five-hour limit, plus $200 a month in API credits |
| Team, standard seat | $25 per seat a month, or $20 billed yearly; 2 to 150 people | More than Pro |
| Team, premium seat | $125 per seat a month, or $100 billed yearly | 5 times a standard seat |
| Enterprise | $20 per seat a month billed yearly, 20 seats minimum, plus usage at API rates | Pay for what you use |
Source: Claude: Plans and pricing. Anthropic adds that "Prices shown don't include applicable tax."
On the API, Claude Fable 5.2, Anthropic's top model, costs $8 per million input tokens and $40 per million output tokens; Claude Sonnet 5.5 costs $2 and $10; Claude Haiku 5.5 costs $0.10 and $0.50 for prompts up to 100,000 tokens; Claude Opus 5.5 costs $4 and $20 (Claude API Docs: Pricing).
What changed in 2026
- Claude Max and Ultra now include monthly API credits ($100 and $200) and Fable, Anthropic's top model, for up to half the weekly limit. Pro pays for Fable with usage credits.
- OpenAI renamed Priority processing to Fast mode on July 30, 2026. Ultrafast on GPT-6.2 Astra and Sol costs 3 times the Standard rate, rising to 6 times on March 1, 2027.
Is a subscription or pay-as-you-go cheaper?
For a person doing daily work, a subscription is usually cheaper and always more predictable. Token billing only wins when the volume is small or the work runs without a person.
Three things move the per-token bill:
- Model size. The top models cost about 80 times more per token than the smallest ones on both price lists ($8 versus $0.10 per million input tokens). Use the small model for sorting, tagging and short replies.
- Batch. Anthropic gives "a 50 percent discount on both input and output tokens" for jobs that can wait (Claude API Docs: Pricing). OpenAI's batch prices are also half its standard rates (OpenAI API Docs: Pricing).
- Token count. The same text can produce more tokens on newer models. Anthropic says its newer tokenizer "produces approximately 30% more tokens for the same text" (Claude API Docs: Pricing). Compare cost per finished task, not price per token.
When to upgrade your plan
Upgrade when limits stop your work, not when a bigger plan looks impressive.
- Start one person on the $20 plan of the tool your team already likes.
- Track for a month how often you hit the five-hour or weekly limit.
- Hitting it a few times a month: buy credits when it happens.
- Hitting it every week: move to the next tier ($100 Max on Claude, $100 Pro on ChatGPT).
- More than one person: a Team or Business plan keeps company data and billing in one account.
Example
A two-person cleaning company
Both owners use Claude Pro for quotes, emails and social posts: $40 a month. They hit limits twice a month. They stay on Pro rather than jump to $200 for two Max seats.
A website chatbot that answers 3,000 questions a month
Each answer reads about 2,000 tokens and writes about 300. On a small model at $0.10 input and $0.50 output per million tokens, that is about $0.60 + $0.45 = $1.05 a month in tokens. Per-token billing is cheap here because no person sits at a keyboard.
A bookkeeper who automates a nightly report with Codex
On ChatGPT Plus ($20 a month) the job keeps hitting the five-hour cap. Moving to Pro at $100 removes the five-hour limit; running it by API key instead makes sense only if the token bill stays under that $100.
FAQ
Is Codex free?
Codex is included in every ChatGPT plan, including Free, but the Free and Go plans have limited Codex use. Heavy use needs Plus, Pro or a business plan.
Does Claude Pro include Claude Code?
Yes. Claude Code is included on Pro, Max, Ultra, Team and Enterprise, but not on Free.
What happens when I hit my usage limit?
You wait for the limit to reset, buy extra credits where the plan offers them, or upgrade. Anthropic notes that how fast you reach a limit depends on the model, chat length and file size.
Are these prices in Canadian dollars?
No. Both companies list prices in US dollars. Anthropic's API docs say "All prices are in USD", and its plan page says tax is extra.
Do ChatGPT Pro's API credits roll over?
No. OpenAI says "Unused credits expire at the end of the applicable billing cycle." Use them each month or lose them.
Different for you?
- You want AI that does marketing work, not a chat window: see AI agents for small business marketing.
- You are not sure what an agent is versus a chatbot: see agentic AI explained.
- You are pricing an AI SEO tool or service rather than a model: see AI SEO agent pricing.
- You want everyday marketing uses for ChatGPT: see ChatGPT for marketing.