🔍 Read the full analysis: The AI Price War: A Guide To OpenAI And Anthropic’s Customer Deals on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
OpenAI and Anthropic have reduced prices for several AI models in recent months, while a report says OpenAI is offering discounts to attract customers using Anthropic. Reported deals include a discount above 30% for Replit and a projected 80% to 90% cost reduction cited by DataCamp’s CEO, which is not a 90% list-price discount. Seat minimums, usage limits and expiring credits also affect what customers pay.
OpenAI and Anthropic have cut prices on several AI models since June, and OpenAI is offering additional discounts to attract customers from Anthropic, according to an October 24 report by The Information. The reported deals range from a discount of more than 30% for Replit to projected total cost savings of 80% to 90% cited by DataCamp’s CEO — figures that describe different things and should not be confused with a single across-the-board price cut.
Vendor pricing pages checked on October 26 show OpenAI’s GPT-6.1 Sol at $2 per million input tokens and $10 per million output tokens, compared with $5 and $30 for the earlier Sol model in June. Its Luna model fell from $1 and $6 to $0.10 and $0.50. These prices reflect different model releases, not necessarily a reduction in the cost of running an identical model.
Anthropic’s Opus pricing also fell across releases: Opus 5 was listed at $5 per million input tokens and $25 per million output tokens in July, while Opus 5.5 was listed at $4 and $20. Anthropic also cut prices for some cached-token reads. Its statements that newer models cost less on typical or agent-heavy workloads are company estimates, not independent measurements of every customer’s bills.
The Information reported that OpenAI is using customer-specific pricing to compete for accounts. Replit was offered a discount of more than 30% on a contract valued at more than $100 million a year, according to a person cited in the report. DataCamp CEO Jonathan Cornelissen said OpenAI’s offer could reduce his company’s costs by 80% to 90%. That is an expected saving against its existing setup, not a discount of that size off OpenAI’s published prices.
Infographic 01 · The price slide
Four months, a fraction of the price.
What a million output tokens cost on each vendor’s price list, from the June previews to today. Bars show the price; shorter is cheaper.
OpenAI · output price per 1M tokens
List prices, standard tier
Anthropic · output price per 1M tokens
List prices, standard tier
Same scale in both panels. Fable 5.1 kept its $50 output price and cut cache reads instead.
The quieter cut: cache reads
matters most for agents
Cache-read price per 1M tokens, same scale in all three.
Woolf estimates cached tokens make up 96–99% of token spend in agentic coding. Anthropic says Opus 5.5 costs 40% less than Opus 5 and Fable 5.1 about 25% (up to about 45%) less than Fable 5 on typical workloads; those are vendor claims. Most rows compare a successor model with its predecessor. Anthropic’s newer tokenizer produces about 30% more tokens for the same text, so per-token prices are not directly comparable across vendors.
Sources: OpenAI API pricing (26 Oct 2026); OpenAI announcements (26 Jun–29 Sep 2026); Anthropic API pricing (26 Oct 2026); Anthropic announcements (30 Jun–22 Sep 2026); Vercel AI Gateway index (Oct 2026); Max Woolf (25 Oct 2026).
Powered by Thorsten Meyer AI
Lower Prices, More Contract Leverage
The price cuts give companies more room to renegotiate contracts and test competing providers. A lower token price can affect the economics of coding assistants and other workloads that make large numbers of model calls. But the relevant figure for a buyer is the full bill: model quality, usage volume, cached-token charges, support and contract terms all affect the cost of a particular task.
There are signs that customer usage can shift quickly. Rippling CEO Parker Conrad said OpenAI’s share of the company’s coding work rose from 16% in May to 52%, while Anthropic’s share fell from 57% to 18%. That is one executive’s account of one company’s use, not evidence of a market-wide shift. Still, it illustrates why pricing has become a competitive tool as well as a procurement issue.
Customers also need to check what is included. A lower headline price may come with changes to usage allowances or seat rules. Those details can determine whether a deal works for a large enterprise, a small business or an individual subscriber.
Popular right now
LG C2 Series 65-Inch OLED evo 4K Smart TV (OLED65C2PUA, 2022)
Ninja CREAMi 7-in-1 Ice Cream Maker, NC301
De'Longhi Linea Classic Espresso Machine with Adjustable Milk Frother
As an affiliate, we earn on qualifying purchases.
A Four-Month Pricing Shift
The reductions followed a series of model releases and pricing changes between June and September 2026. Data scientist Max Woolf tracked the changes in an October 25 post; the source material says his figures were checked against vendor announcements and pricing pages retrieved October 26. Anthropic set an introductory price for Claude Sonnet 5 permanently in August, while OpenAI reduced prices for Luna and later introduced lower-priced GPT-6 models.
Some savings are tied to cached input tokens, rather than a lower standard input or output rate. Coding agents can reuse conversation history across turns, making cached-token pricing relevant to some workloads. Woolf estimates that cached tokens make up 96% to 99% of token spending in agentic coding; this is his estimate, not a vendor-wide audited figure. Anthropic has said its Fable 5.1 model can cost less on typical and highly agentic workloads, but those comparisons depend on the workload and the company’s stated methodology.
Reported customer terms also vary. The source material says Anthropic added capped, expiring API credits to its Max and Team plans and that its Enterprise plan has a 20-seat minimum. It also says OpenAI reduced usage included with new $200 Pro subscriptions without changing the monthly price. These changes show why list prices alone cannot describe the terms a buyer receives.
“I think ChatGPT is better now.”
— Amjad Masad, Replit CEO, speaking at TechCrunch Disrupt on October 16
Deal Terms Vary by Customer
The reported discounts do not establish a standard offer available to every buyer. The Information’s article is paywalled, and the source material relies in part on its public summary and subsequent coverage. The full contracts, eligibility rules and comparison periods for the reported offers are not public in the material provided.
It is also unclear how much of the reported savings comes from lower rates, different usage patterns, bundled seats or other contract terms. Replit’s relationship with OpenAI is described differently across accounts: Masad called OpenAI its preferred partner while saying the company remained model agnostic, and a person familiar with the partnership told TechCrunch it was still in place. The underlying contract terms and the exact scope of its use are not stated.
Price comparisons between model generations do not show whether a newer model performs as well on each customer’s tasks. Buyers would need to test their own workloads and account for the usage limits and credit rules in their contracts. The source material also ends mid-sentence while describing a bank seat offer, so the complete terms of that deal are unavailable here.
Watch Renewals and Usage Limits
Companies evaluating AI contracts can compare published rates with customer-specific offers, test models on their own workloads and review the total cost across token use, seats, credits and support. Existing customers may use competing bids and usage data in renewal discussions, but the source material does not identify a single industry-wide deadline or forthcoming pricing change.
For subscribers, the next changes to watch are usage allowances and plan terms, not only monthly fees. Vendors may revise included usage or credit rules while keeping a subscription price unchanged. Further details about the reported customer deals, including their duration and conditions, would clarify how broadly the discounts are available.
Key Questions
Did OpenAI cut its prices by 90%?
No broad 90% list-price cut is described. DataCamp CEO Jonathan Cornelissen said the company expected its total costs to fall by 80% to 90% under an OpenAI offer, compared with its current setup. A separate reported Replit offer was a discount of more than 30%.
How much did the listed model prices change?
The source material says OpenAI’s Sol price moved from $5 per million input tokens and $30 per million output tokens in June to $2 and $10 for GPT-6.1 Sol in September. Luna moved from $1 and $6 to $0.10 and $0.50. The figures compare different model releases.
Are customers leaving Anthropic for OpenAI?
The report describes OpenAI courting Anthropic customers, and Rippling’s CEO said his company’s coding-work share shifted toward OpenAI. Replit also named OpenAI its preferred partner, while saying it remained model agnostic. These examples do not establish a market-wide customer movement.
What should AI buyers check beyond the token price?
Buyers should compare performance on their own tasks and check usage caps, cached-token rates, seat minimums, credit expiry and included allowances. Those terms can change the total cost even when a published model rate or monthly subscription price looks low.
Source: ThorstenMeyerAI.com
