OpenAI Pricing Plans: Every Plan and Price, Explained (2026)
Most guides to OpenAI pricing plans quietly make the same mistake: they mash a “$8–$200 a month” subscription price and a “pay-per-token” API rate into one table, as if they were the same bill. They’re not. OpenAI runs two separate pricing lines — a per-seat monthly ChatGPT subscription and a pay-per-token API — and knowing which one you’re on is the whole game.
Here’s the short version, checked against OpenAI’s own pages on July 23, 2026:
- ChatGPT subscription plans — billed per seat, per month, from $0 (Free) up to the $100+ Pro tier for individuals, and $20–$25 per user for Business. See OpenAI’s ChatGPT pricing page.
- OpenAI API — billed per token (or per second, per minute, per 1,000 calls) based on usage, with no subscription at all. Flagship text models run roughly $1–$30 per million tokens depending on the model, per OpenAI’s API pricing page.
- Separately, OpenAI’s image models can be reached per image through a third-party aggregation layer — a repackaging of the API’s image endpoints for convenience, not a third pricing plan OpenAI itself offers.
This guide lays out both OpenAI lines, plan by plan and model by model, covers that third-party per-image option briefly, and then answers the question the price tables don’t: which one should you actually pay for. OpenAI adjusts these numbers often, so treat every figure here as a dated snapshot — confirm against the live pages before you commit.
The two lines of OpenAI pricing plans (and why they’re not the same)
The single most common error in OpenAI pricing write-ups is treating “OpenAI” as one product with one price. It isn’t. OpenAI itself sells two products on two billing systems:
| Pricing line | What you’re buying | Unit | Who it’s for |
|---|---|---|---|
| ChatGPT subscription | Access to the ChatGPT app (web/desktop/mobile) | $ per seat, per month | People using the chat product |
| OpenAI API | Programmatic model access to build your own product | $ per 1M tokens (also $/sec, $/min, $/1k calls) | Developers |
The practical upshot: a ChatGPT subscription gives you no API credits, and API spending gives you no ChatGPT app access — they’re billed and topped up entirely separately. And when someone quotes “OpenAI costs $0.02,” you can’t tell what they mean until you know whether that’s per token or a rounding of a monthly seat. Comparing the raw numbers across these units is a category error. So the rest of this article keeps them apart. (There’s also a third-party per-image way to reach OpenAI’s image models, which isn’t an OpenAI billing line at all — covered near the end.)
ChatGPT subscription plans
The ChatGPT side has six tiers — four aimed at individuals, two at organizations. Prices are per user, per month, and come straight from OpenAI’s ChatGPT pricing page.
| Plan | Price | Best for |
|---|---|---|
| Free | $0 | Trying ChatGPT; limited GPT-5.5 Instant, limited image generation and deep research |
| Go | $8 / month | Light users who want more messages and uploads (note: this plan may include ads) |
| Plus | $20 / month | Regular users who want advanced GPT-5.6 reasoning, projects, and expanded limits |
| Pro | From $100 / month | Power users: $100 for 5× Plus usage, with a higher-priced 20× tier, plus the Pro reasoning tier |
| Business | $20 / user / month (annual) · $25 (monthly) | Teams (2+ seats): shared workspace, SSO, admin controls, no training on your data |
| Enterprise | Custom | Large deployments; contact sales for pricing |
ChatGPT’s own pricing page, captured July 23, 2026 — Free at $0, Go at $8, Plus at $20, and Pro starting “From $100” per month, which confirms Pro’s entry price is $100, not the $200 figure many older pages quote.
A few things the older ranked pages get wrong or skip:
- Pro starts at $100, not $200. The page shows “From $100/month” and describes a usage toggle: $100/month buys 5× Plus usage, with a higher-priced tier for 20×. If you’ve only ever seen Pro quoted at $200, that’s the top of the toggle, not the entry point.
- “Team” is now “Business.” OpenAI folded the old Team plan into Business, at $20 per user/month billed annually (2-seat minimum) or $25 per user/month billed monthly.
- Enterprise pricing is genuinely custom. OpenAI lists “Custom pricing — contact our sales team.” Third-party per-seat estimates floating around are not official numbers, so budget from a sales conversation, not a blog.
There are also two programs worth knowing if you qualify: ChatGPT for Teachers (free for verified U.S. K-12 educators through June 2027) and up to 75% off ChatGPT Business or Enterprise for nonprofits.
OpenAI API pricing (pay per token)
The API is the other line entirely. There’s no monthly seat — you’re billed for what you process, mostly in tokens (roughly 3–4 characters each). Every model has an input rate (what you send), an output rate (what it generates), and usually a discounted cached input rate (for repeated context). Prices below are per 1 million tokens, at the Standard tier, short-context, from OpenAI’s API pricing page.
Flagship text and reasoning models:
| Model | Input | Cached input | Output |
|---|---|---|---|
| gpt-5.6-sol | $5.00 | $0.50 | $30.00 |
| gpt-5.6-terra | $2.50 | $0.25 | $15.00 |
| gpt-5.6-luna | $1.00 | $0.10 | $6.00 |
| gpt-5.5 | $5.00 | $0.50 | $30.00 |
| gpt-5.5-pro | $30.00 | — | $180.00 |
| gpt-5.4 | $2.50 | $0.25 | $15.00 |
| gpt-5.4-nano | $0.20 | $0.02 | $1.25 |
| gpt-5.4-pro | $30.00 | — | $180.00 |
OpenAI’s API pricing page, captured July 23, 2026 — flagship gpt-5.6-sol at $5.00 input / $30.00 output and gpt-5.5-pro at $30.00 / $180.00 per 1M tokens (short context), with the March 5, 2026 regional 10% uplift noted directly below the table.
The spread is wide: the smallest model (gpt-5.4-nano) is about 25× cheaper on input than the flagship gpt-5.6-sol, and the -pro reasoning models cost roughly 6× the standard flagship. Model choice, not haggling, is the lever that moves an API bill. (Long-context requests are priced higher — gpt-5.6-sol roughly doubles to $10 input / $45 output beyond the short-context window.)
The API also meters non-text work in its own units:
| Category | Model (example) | Price |
|---|---|---|
| Image (token-metered) | gpt-image-2 | $8.00 input / $30.00 output per 1M tokens |
| Realtime audio | gpt-realtime-2.1 | $32 input / $64 output per 1M audio tokens |
| Transcription | gpt-4o-transcribe | $0.006 / minute |
| Video | sora-2 / sora-2-pro | $0.10 / sec (720p) up to $0.70 / sec (1080p) |
| Web search tool | all models | $10 / 1,000 calls |
Two billing details that catch people out: models released on or after March 5, 2026 add a 10% uplift on regional data-residency endpoints, and cheaper tiers (Batch, Flex) can roughly halve the Standard rate for workloads that tolerate latency.
On the image models specifically: the table above shows the raw token rate, but for image work most people want a per-image number. That’s a topic in its own right — the low/medium/high per-image ladder, which model to build on, and the DALL·E retirement — so we’ve kept it in a dedicated guide: see the OpenAI image generation API breakdown. Here it’s enough to know the image models are billed by token on the direct API, and that there’s a simpler per-image way to reach them, covered next.
Which OpenAI plan should you pick?
Price tables don’t answer the real question. Here’s the decision, by what you’re doing:
- You want to use ChatGPT → subscribe. Casual use is fine on Free or Go ($8); Plus ($20) is the default for regular professional use; Pro (from $100) only pays off if you’re hitting Plus limits or need the top reasoning tier. For a team, Business ($20–$25/user) adds admin and privacy controls; Enterprise is for scale and custom terms.
- You want to build on OpenAI’s models → use the API and budget per token. Start from the cheapest model that clears your quality bar (
gpt-5.4-nanoorgpt-5.6-lunabeforegpt-5.6-sol), lean on cached input, and consider Batch/Flex for non-urgent jobs. - You only need OpenAI’s image models → you don’t have to touch either meter head-on. Those models can be reached through an aggregation layer that prices them per image.
On that last option: Modellix carries four OpenAI image models — gpt-image-2 and gpt-image-1.5 plus their edit variants — metered per image ($0.0041–$0.3943 for gpt-image-2, $0.0138–$0.2334 for gpt-image-1.5), pay-as-you-go with no monthly fee, and a per-call log of cost and latency on every request. Two honest limits keep this in perspective: it’s a distribution layer for OpenAI’s models, not OpenAI itself and not a cheaper or exclusive source; and it carries only those image models — not the text, audio, Sora, or gpt-image-1/-mini models — so for anything beyond OpenAI image generation you’d go to the direct API. What a per-image layer buys you is a simpler unit to reason about when images are all you need.
OpenAI pricing plans FAQ
Is ChatGPT free?
Yes — the Free plan is $0 and available to everyone, with limited access to GPT-5.5 Instant and capped image generation and deep research. Paid plans start at $8/month (Go).
How much is ChatGPT Plus vs Pro?
Plus is $20/month. Pro is “from $100/month” — $100 buys 5× Plus usage, with a higher-priced 20× tier — plus Pro’s higher reasoning tier. Pro only makes sense if you’re regularly hitting Plus limits.
Does a ChatGPT subscription include API access?
No. ChatGPT subscriptions and the OpenAI API are billed separately. A Plus or Pro plan gives you the app, not API credits; API usage is paid per token and does not include ChatGPT app access.
How is OpenAI API pricing calculated?
By usage, mostly per token, split into input, output, and (discounted) cached input, priced per million tokens. Flagship text models run about $1–$30 per 1M tokens; audio, video, and transcription are metered per token, per second, or per minute.
How much does ChatGPT Business or Enterprise cost?
Business is $20 per user/month billed annually (2-seat minimum) or $25 per user/month billed monthly. Enterprise is custom pricing — you contact OpenAI’s sales team.
What’s the cheapest way to use OpenAI’s image models?
On the direct API they’re token-metered; per image, gpt-image-2 runs roughly $0.006–$0.211 at 1024×1024 depending on quality. Reaching them per image through a third-party aggregation layer can be simpler for image-only work — you can browse the OpenAI options among Modellix’s text-to-image models — but it’s a repackaging of the same API, not a discount. See the OpenAI image generation API guide for the full per-image breakdown.
Plan names and pricing reflect OpenAI’s public pages as of July 2026 and change frequently — validate against OpenAI’s live ChatGPT and API pricing before you commit. Reach OpenAI’s image models, plus other image and video models, through a single API key at modellix.ai.
