The Direct Answer: Per Seat, Annual, Negotiated
ChatGPT Enterprise is priced per seat, billed annually, and negotiated on every contract — OpenAI does not publish a sticker price. The working benchmark is around $60 per user per month, with negotiated deals landing anywhere from roughly $45 to $75 depending on seat count, term and usage profile. Because there is no list price, the rate you pay is almost entirely a function of how you buy: the same 500-seat deployment can sit five or more dollars per seat apart on negotiation quality alone.
That makes preparation the whole game. OpenAI Enterprise is one component of a wider AI estate, and it should be priced alongside the API and any platform commitments rather than in isolation — the broader framework sits in our complete guide to AI procurement.
The Pricing Structure and the Floor
Three structural rules set the floor. First, the 150-seat minimum: OpenAI will not sell ChatGPT Enterprise below it. Second, the annual commitment, billed as a prepaid invoice — there is no month-to-month enterprise option. Third, the per-seat model itself, which is a flat charge regardless of how heavily each user engages. At the $60 benchmark, 150 seats is about $108,000 per year; at 500 seats and a negotiated $50, roughly $300,000 per year.
Organisations below the 150-seat floor are steered to ChatGPT Business (the Team tier), which is sold per seat without the enterprise minimum but omits the enterprise controls, SLAs and commercial protections. The choice between seat-based and consumption-based models is itself a design decision — we work it through in AI seat-based vs consumption pricing.
Volume Discount Tiers
Volume and term are the primary discount drivers. The pattern in 2025–2026 enterprise deals:
| Seat Band | Typical Discount vs Benchmark | What Opens Up |
|---|---|---|
| 150–299 | 0–10% | Standard enterprise terms; limited movement |
| 300–499 | 10–15% | Free pilot cohort; price-cap negotiation |
| 500–999 | 15–25% | Blended rates; bundled API credits |
| 1,000+ | 25–40% | Custom structures; multi-year blended deals |
A two- or three-year commitment adds a further 5–15% on top of the volume band. Buyers committing meaningful seats routinely land 25–40% below the $60 benchmark — but the deepest discounts come bundled with term length, so model the flexibility you are trading away before signing multi-year.
The Levers Beyond Per-Seat Price
Per-seat price is the headline, but the value sits in the surrounding terms. The levers that consistently move an OpenAI Enterprise deal: a free pilot cohort (typically 10–25 seats for 60–90 days) to prove adoption before you commit at scale; bundled API credits folded into the seat licence so your developer usage rides the same contract; an annual price-increase cap (2–3% is the market standard) to stop renewal surprises; onboarding and professional-services credits; and commercial protections — data isolation, IP-ownership confirmation and SLA upgrades — that are absent from the standard self-serve terms.
These protections matter as much as the price. The contractual red flags to watch are catalogued in the AI procurement checklist, and the deployment decision between OpenAI direct and Azure OpenAI — which changes both pricing and data terms — is compared in Azure OpenAI vs OpenAI direct cost. Our AI procurement advisory runs these negotiations end to end.
The Seats-Plus-Tokens Trap
The most common costing error is treating ChatGPT Enterprise seats and OpenAI API tokens as separate purchases. Seats are a per-seat application licence; the API is billed per token and invoiced separately. Most enterprises run both — seats for end-user productivity, tokens for products and automation — and buying them on two disconnected contracts forfeits the largest lever of all: trading API committed spend against seat pricing inside a single negotiation.
Token pricing is itself more negotiable than most buyers assume, as we set out in are AI token prices negotiable. Align both to the same commercial calendar — the timing logic in planning your IT contract renewal calendar applies to AI contracts too. When an OpenAI commitment is material, request a confidential briefing — we benchmark the per-seat rate and structure seats and tokens as one deal.