Short answer
If by “GPT” you mean ChatGPT, there’s a free plan; paid individual plans range from $8 to $500 per month in the U.S. If you mean using GPT in your own app or production workflow, API usage is billed separately, typically by tokens processed. At the listed standard short-context API rates, 1 million input tokens plus 200,000 output tokens would cost an estimated $0.20 on GPT-6 Luna, $4 on GPT-6.1 Sol, or $20 on GPT-6 Astra—but short-context eligibility depends on a per-request rule that is not confirmed here, so these figures are conditional examples, not a production quote. Check the OpenAI API pricing page for the applicable rate.
U.S. plan prices shown below are in U.S. dollars. ChatGPT prices can vary by country and currency; Go has localized pricing in some markets. API prices below are in U.S. dollars per million tokens. Confirm the price and plan shown for your account before purchasing.
ChatGPT plan prices
| Plan | Price | Billing unit and important qualification |
|---|---|---|
| Free | $0 | Individual ChatGPT access; tools and usage limits vary. |
| Go | $8/month in the U.S. | Monthly subscription; price is localized in some markets. |
| Plus | $20/month | Per account, billed monthly. API usage is not included. |
| Pro 100 | $100/month | Per account, billed monthly. |
| Pro 200 | $200/month | Per account, billed monthly. |
| Pro 500 | $500/month | Per account, billed monthly; the Pro tier that includes Ultrafast. |
| Business Standard | $20/user/month annually, or $25/user/month monthly | Per seat. The general requirement is at least two paid seats; confirm the terms shown for your workspace. |
| Business Premium | $100/user/month annually, or $125/user/month monthly | Per seat; higher Work and Codex usage than Standard. The general requirement is at least two paid seats; confirm the terms shown for your workspace. |
| Enterprise | Custom quote | Contact OpenAI Sales; there is no public fixed per-seat price in the cited information. |
OpenAI says Pro comes in three monthly tiers; Pro 200 includes more usage than Pro 100, and Pro 500 has the highest included usage and Ultrafast. Business Standard and Premium prices depend on billing commitment. The exact plan and limits that fit depend on how you use ChatGPT, not just the headline price.
GPT API pricing for production use
The API is a separate product and billing system from ChatGPT subscriptions. API charges depend on the model and features used, including token usage and applicable tool or storage charges. The table shows standard, short-context rates in U.S. dollars per million tokens; it is not a monthly subscription fee.
| API model | Input per 1M tokens | Cached input per 1M tokens | Output per 1M tokens |
|---|---|---|---|
gpt-6-luna |
$0.10 | $0.01 | $0.50 |
gpt-6.1-sol |
$2.00 | $0.10 | $10.00 |
gpt-6-astra |
$10.00 | $1.00 | $50.00 |
These are standard rates for the short-context tier shown on OpenAI’s API pricing page. The specific token threshold and per-request rule that determine short- versus long-context pricing are not established here. Treat the figures as conditional examples, not as confirmed rates for a long-context request. Long-context and different processing modes have different rates. Cached-input pricing applies only when tokens qualify as cached; tools and other features may also add charges. Check the pricing page for the rate that applies to your model, context, and processing choice.
Example: estimate a production workload
Suppose your requests in aggregate use 1 million input tokens and 200,000 output tokens, and each request qualifies for the standard short-context tier. The exact eligibility threshold and trigger are not confirmed here, so do not apply this estimate to requests whose context may qualify for long-context pricing. With no cached tokens or additional tool charges:
- GPT-6 Luna: (1 × $0.10) + (0.2 × $0.50) = $0.20
- GPT-6.1 Sol: (1 × $2) + (0.2 × $10) = $4
- GPT-6 Astra: (1 × $10) + (0.2 × $50) = $20
This is an arithmetic estimate, not a quote or a claim that the models perform equally on your task. Actual usage, context tier, caching, and tools can change the bill. For production budgeting, measure token use on representative requests and confirm the applicable context tier and rate for the exact model and processing option you plan to deploy.
Which option should you choose?
- You want to use ChatGPT yourself: Start with Free if its available access is enough. Consider Go or Plus for a paid individual plan; compare current features and limits in your account before upgrading. Plus is $20/month, while Go is $8/month in the U.S.
- You repeatedly need more included ChatGPT usage: Compare the Pro tiers by price and the allowance or speed feature you actually need. Pro costs $100, $200, or $500 per month; Ultrafast is included only with Pro 500.
- A team needs a shared workspace: Compare Business seat costs against individual plans, including the annual commitment and general two-paid-seat requirement. Enterprise is a separate, custom-quoted option for organizations that need a sales agreement.
- You’re building an app or running an automated production workflow: Budget using API rates rather than assuming a ChatGPT subscription covers it. Shortlist models, then test them on representative inputs for output quality, latency, and cost before choosing. No single model is automatically the best value for every workload.
Bottom line: ChatGPT starts at $0, Plus is $20/month, and Pro ranges from $100 to $500/month. For a production API, costs vary with model and usage; estimate from the input and output tokens your workload actually processes, and confirm which context rate applies. ChatGPT subscriptions and API billing are separate.