If you're evaluating ChatGPT for your business, OpenAI's pricing page is only the starting point. The published seat price is easy enough to understand. Forecasting what your organization will actually spend is much harder.
In SpendHound's 2026 AI Spend Report, 46% of finance and procurement leaders said they exceeded their AI budget, while 57% said they either weren't confident they were paying a fair price for AI tools or simply didn't know. That's not surprising. Enterprise AI pricing has become increasingly difficult to forecast as organizations combine ChatGPT Business or Enterprise with API usage, reasoning models, Codex, and rapidly expanding adoption across teams.
This guide combines OpenAI's published pricing with SpendHound's pricing data from 1,000+ organizations and practical guidance from procurement leaders who negotiate AI contracts every week.
Below, you'll find how ChatGPT pricing works across every plan, what organizations are actually spending, which pricing levers matter during enterprise negotiations, and how to approach your next renewal with benchmark-backed pricing instead of vendor guidance alone.
ChatGPT Business costs $20 to $25 per user per month, depending on whether you pay annually or monthly. ChatGPT Enterprise uses custom pricing that's negotiated directly with OpenAI. Organizations building with OpenAI also pay separately for API usage, which is billed based on token consumption and model selection.
Individual users can also choose Free, Go, Plus, and Pro plans, but most finance and procurement teams evaluating ChatGPT will focus on Business, Enterprise, and the API.
OpenAI offers four plans for individual users. Free includes limited access to ChatGPT, while Go ($8/month), Plus ($20/month), and Pro ($100 or $200/month) increase usage limits, expand access to advanced models, and unlock additional features. Unlike Business, these plans are designed for individual subscriptions rather than centralized company management, and none of the Pro tiers offer annual pricing.
For finance and procurement teams, these plans usually matter only at the margins. Most organizations standardize on ChatGPT Business or Enterprise, then reserve Pro for a small number of power users whose workloads justify the additional cost. Giving every employee the highest-tier plan rarely makes financial sense and can increase AI spend without delivering proportional value.
Business costs $20 per user per month with annual billing or $25 per user per month when billed monthly, with a two-user minimum. OpenAI offers two Business seat types: standard ChatGPT seats and Codex-only seats for engineering teams. Before rolling out licenses broadly, it's worth deciding which users actually need Codex. A company-wide upgrade is rarely necessary and can increase costs without delivering much additional value.
ChatGPT Enterprise is OpenAI's custom offering for larger organizations that need centralized administration, advanced security controls, and enterprise support. Unlike ChatGPT Business, ChatGPT Enterprise doesn't have published pricing. Every agreement is negotiated directly with OpenAI based on the organization's size, expected usage, deployment requirements, and commercial terms.
As a result, there isn't a standard Enterprise price. Two companies with similar employee counts can end up with very different contracts depending on how they plan to use ChatGPT.
Unlike Business, Enterprise pricing isn't built around a published per-user subscription. OpenAI works with customers to structure contracts around their deployment, which may include user licenses, committed usage, API consumption, or a combination of the three.
For many organizations, API usage becomes an important part of the commercial discussion. Many organizations deploy ChatGPT Enterprise for employees while building internal applications on OpenAI's API, creating two separate sources of AI spend that should be evaluated together.
Several factors influence what an organization ultimately pays:
Because Enterprise contracts combine multiple pricing components, organizations with similar headcounts can have substantially different annual spend.
OpenAI doesn't publish typical Enterprise contract values, making it difficult to understand what organizations actually spend.
Based on aggregated spend data from SpendHound's vendor dataset, average annual OpenAI spend (as of July 2026) is:
These figures represent observed average annual spend across organizations in SpendHound's dataset rather than OpenAI's published pricing. Individual contracts vary considerably based on deployment size, API usage, and negotiated commercial terms, but they illustrate how quickly enterprise AI costs can grow beyond the price of individual user licenses.
Average contract values provide useful context, but they don't tell you whether your agreement is competitive. OpenAI Enterprise pricing varies significantly based on company size, deployment model, expected usage, and commercial terms, so two organizations can receive very different proposals.
That's why procurement teams rely on peer-based pricing benchmarks rather than market averages alone. SpendHound compares your pricing against organizations with similar characteristics, including company size, deployment profile, and contract structure. That provides a more relevant view of what comparable companies are paying before you enter negotiations.
Example of a SpendHound OpenAI pricing benchmark showing how a company's pricing compares with similar organizations.
Organizations building applications with OpenAI pay separately for API usage. Unlike ChatGPT Business or Enterprise, which are primarily licensed by user, the API is billed based on token consumption. Every request includes input tokens sent to the model and output tokens generated in response. More capable models cost more per token, making model selection one of the biggest drivers of API spend.
The model you choose is only part of the story. Total API spend is driven by how often applications call the API, how much context each request includes, and how broadly AI is embedded across your products and internal workflows. Two organizations can use the same model yet end up with dramatically different bills because one processes far more requests or sends significantly more context with each prompt.
OpenAI also offers several ways to reduce API costs. Cached inputs are billed at a steep discount when applications repeatedly reuse the same prompts or reference material. The Batch API cuts costs for asynchronous workloads that don't require real-time responses. For organizations running high-volume AI applications, these optimizations can meaningfully reduce annual spend without changing the underlying models.
Codex is included with ChatGPT Plus, Business, and Enterprise plans, subject to plan-specific usage limits. Organizations can also provision Codex-only seats for engineering teams, allowing developers to use AI coding tools without assigning a full ChatGPT Business license.
For many companies, engineering becomes one of the fastest-growing sources of AI spend. AI coding tools tend to see rapid adoption once they're rolled out, and usage often grows much faster than initial forecasts. When budgeting for OpenAI, finance and procurement teams should evaluate expected software engineering usage alongside general ChatGPT adoption rather than treating developer AI spend as a separate initiative.
The right plan depends less on company size than how your organization plans to use AI.
Choose Business if you have roughly 5 to 150 users, most usage is chat and Projects rather than custom applications, and predictable per-seat pricing matters more than deep governance controls. Plan for API access separately once engineering starts calling OpenAI's models programmatically, since that usage won't route through your Business seats.
Choose Enterprise when security and governance requirements, not headcount, become the deciding factor: SCIM, EKM, RBAC, data residency, and IP allowlisting are procurement requirements your security or compliance team is asking for, not features you'd like to have.
Choose API-only when the primary use case is engineering or product teams building agents or internal tools, where usage is tied to applications rather than individual seats.
For most midmarket organizations and up, the real answer is a hybrid: Business seats for the broad workforce, API access for engineering. That's not a compromise, it's what usage patterns actually look like once a company gets past a few dozen employees.
OpenAI remains one of the most widely adopted enterprise AI platforms. Across SpendHound's Vendor Trends dataset, the number of organizations with OpenAI spend increased from approximately 1,000 in June 2025 to 1,100 by May 2026 before leveling off, suggesting that ChatGPT has already reached broad adoption across the enterprise market.
OpenAI maintained a large enterprise footprint throughout the past year, but Anthropic closed the gap quickly. Between June 2025 and May 2026, the number of organizations with Anthropic spend increased from 478 to 1,100, while OpenAI remained relatively stable at approximately 1,100 observed customers. By May 2026, both vendors appeared in roughly the same number of organizations across SpendHound's Vendor Trends dataset, suggesting that many organizations now deploy both ChatGPT and Claude rather than standardizing on a single provider.
For finance and procurement teams, that changes how AI budgets are managed. Instead of evaluating one strategic AI vendor, many organizations are balancing multiple model providers across different use cases. Engineering may prefer one model for coding, while legal, finance, or operations rely on another. As a result, forecasting AI spend increasingly means managing a portfolio of vendors rather than a single enterprise contract.
OpenAI Enterprise agreements are more complex than traditional software contracts because they often combine user licenses with usage-based API commitments. Procurement teams should evaluate both pricing models together to understand the organization's total cost of ownership and negotiate the right commercial structure.
Because OpenAI still monetizes ChatGPT Business and Enterprise primarily through user licenses, procurement teams shouldn't expect large API commitments to eliminate seat costs. Instead, use your expected deployment across both licensing models to negotiate the overall commercial structure and forecast costs more accurately.
The following strategies come directly from SpendHound's Procurement Experts, who help customers negotiate OpenAI agreements every week.
The best negotiating leverage exists before your renewal becomes urgent. Once you're inside the final weeks of a contract, your ability to benchmark pricing, evaluate alternatives, or change commercial terms becomes much more limited.
"You may have no intention of leaving OpenAI, but you're in a much stronger negotiating position six months before renewal than you are six weeks before," says Zack Hildenbrandt, Procurement Team Lead at SpendHound. "By then, everyone in the room knows it's much less feasible. Being well ahead and well prepared for those renewals is the most effective lever we've seen."
The AI model market is evolving quickly, and today's best-performing model may not be the right choice a year from now. When negotiating OpenAI agreements, avoid commercial terms that make it unnecessarily difficult to shift workloads as pricing, performance, or business requirements change.
"The biggest thing I'd look to optimize for in terms of AI model spend is flexibility between vendors," Hildenbrandt says. "Models are regularly leapfrogging each other in terms of performance and cost effectiveness, so not locking yourself into one vendor is the most important piece."
Many organizations now deploy multiple AI models across different teams and use cases. Long-term commitments, restrictive minimum spend requirements, or pricing structures that only make sense if you remain with a single provider can reduce that flexibility over time.
Published pricing tells you very little about what enterprise customers actually pay. Going into a negotiation with comparable pricing data gives you a much stronger starting point than relying on list prices or anecdotal conversations.
SpendHound's data provides one useful point of reference: organizations with fewer than 1,000 employees spend an average of $44,318 per year on OpenAI, while organizations with more than 1,000 employees spend an average of $318,506 per year. Those figures illustrate how widely Enterprise deployments can vary, but they're only a starting point.
The strongest negotiating position comes from comparing your pricing and commercial terms with organizations that share a similar company size, deployment profile, and contract structure. SpendHound customers can access those company-specific pricing benchmarks before renewal conversations begin.
User licenses and API commitments may use different pricing models, but they should be part of the same commercial discussion. Looking at both together gives procurement teams a clearer view of total cost of ownership and helps ensure contract terms align with how the organization plans to use OpenAI over the life of the agreement.
To negotiate both effectively, you also need to forecast both. Seat counts are relatively predictable. API usage usually isn't. Before agreeing to annual commitments, work with engineering and business stakeholders to estimate how quickly AI usage is likely to expand over the next year. Forecasting too conservatively can lead to expensive overages, while overcommitting may leave your organization paying for capacity it never uses.
A favorable seat price doesn't guarantee a favorable contract. Usage thresholds and automatic overage pricing can have a much larger impact on total costs as AI adoption grows.
"Even after an AI renewal is finalized, there are often meaningful savings opportunities that get overlooked," Hildenbrandt says. "One of the first things we evaluate is usage tier thresholds and automatic overage pricing, since many customers focus heavily on seat costs during negotiations and underestimate how quickly usage-based charges can grow over time."
Business and Enterprise agreements typically include provisions covering customer data and AI model training, but don't assume every protection discussed during the sales process is reflected in the contract. Verify the language before signing.
Auto-renewal clauses deserve as much attention as price. AI models are evolving quickly, and automatic renewals can leave organizations locked into commercial terms that no longer reflect the market.
"Auto-renewals can lock you into a supplier you'd prefer to move away from, sometimes for multiple additional years if not removed," Hildenbrandt says. "Especially with how rapidly AI tools are evolving, it's important to make sure you don't get locked into a tool that could potentially fall behind the rest of the market."
If the renewal language doesn't work for your organization, push to change it. Notice periods, auto-renewal clauses, and contract length are all negotiable. Jason Edick, Procurement Team Lead at SpendHound, says buyers often have more flexibility than they realize. "You're always able to change contract terms—you just need to take a firm stance and explain why," he says. "Most of the time, they won't ignore the request."
Comparing AI pricing isn't as simple as comparing published seat prices or API rates. The table below summarizes how ChatGPT compares with other leading AI vendors on both dimensions. Microsoft Copilot is included because it's the AI assistant many enterprise buyers already have access to through Microsoft 365.
On published pricing, ChatGPT is broadly competitive with the other leading foundation model providers. Business pricing starts at roughly the same level as Claude Team Standard, while GPT-5.6 Sol is priced similarly to Anthropic's flagship Opus 4.8 model. Google's Gemini continues to offer the lowest published API pricing in this comparison.
Published pricing only tells part of the story. SpendHound's aggregated vendor spend data shows OpenAI customers spending less than Anthropic customers on average across both SMB and Enterprise organizations. OpenAI customers spend an average of $44,318 per year in SMB organizations (50–1,000 employees) and $318,506 per year in Enterprise organizations (1,000+ employees), compared with $55,993 and $441,517, respectively, for Anthropic. The difference is more likely driven by deployment patterns, usage volume, and model selection than by published pricing alone.
ChatGPT spending becomes harder to manage as adoption grows. What starts as a handful of Business seats often expands into Enterprise licenses, API usage, Codex, and AI-powered applications across multiple departments. Without visibility into how that usage is changing, it becomes increasingly difficult to forecast renewals and control spend.
SpendHound helps finance, procurement, and IT teams understand both sides of the equation: what they're spending internally and what comparable organizations are paying externally. Our platform provides OpenAI pricing benchmarks to help you negotiate with confidence, while AI Spend Visibility gives you a centralized view of ChatGPT and OpenAI API usage by model, user, and team so you can forecast costs before they become budget surprises.
Whether you're evaluating ChatGPT for the first time or preparing for an Enterprise renewal, SpendHound can help you understand what you're spending, what comparable organizations are paying, and where you have opportunities to negotiate better commercial terms.
Ready to see how SpendHound can help? Request a demo to see AI Spend Visibility, OpenAI pricing benchmarks, and procurement insights in action.
ChatGPT Business costs $20 per user per month with annual billing or $25 per user per month when billed monthly, with a two-user minimum. Organizations that also use OpenAI's API pay separately based on token usage.
ChatGPT Enterprise is custom-priced by OpenAI's sales team. Observed spend in SpendHound's vendor dataset shows organizations with more than 1,000 employees spend an average of $318,506 per year, although actual costs vary by deployment size, API usage, and commercial terms.
ChatGPT charges both per user and by usage, depending on how it's deployed. ChatGPT Business and Enterprise are licensed per user, while OpenAI's API is billed based on token usage. Organizations may also incur additional usage-based costs through API deployments, Codex, and advanced reasoning models.
Yes. Procurement teams commonly negotiate commercial terms such as volume discounts, usage commitments, auto-renewal language, data handling provisions, and contract structure. Benchmarking comparable agreements before renewal can strengthen your negotiating position.
It depends. Published ChatGPT and Claude pricing is broadly comparable, but total cost depends on deployment size, API usage, and adoption patterns. SpendHound's vendor dataset shows Anthropic customers spend more on average than OpenAI customers, although that difference appears to be driven more by usage than published pricing.
Codex is included with ChatGPT Plus, Business, and Enterprise plans, subject to plan-specific usage limits. Organizations can also purchase Codex-only seats, while the OpenAI API is billed separately based on token consumption for applications, automations, and integrations.
Book a demo below and we'll get you set up with our team.
