🔍 Read the full analysis: How OpenAI And Anthropic Are Trying To Win Over Each Other’s Users 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 is using reported contract discounts to compete for customers of Anthropic, while both companies have cut prices for several models in recent months. Some customers report shifting work toward OpenAI, but headline savings figures do not all represent discounts off list prices, and plan limits and contract terms also affect total costs.
OpenAI is offering discounts to win business from Anthropic, according to reporting published by The Information on October 24, adding a customer-acquisition push to months of falling prices for both companies’ AI models. Some companies say they have shifted more work to OpenAI, but the reported savings vary by customer and should not be confused with a uniform discount from OpenAI’s list prices.
The Information reported that OpenAI offered Replit a discount of more than 30 percent on a contract valued at more than $100 million a year, citing a person familiar with the deal. A separate U.S. bank with more than $7 billion in assets reportedly received 500 ChatGPT seats for the price of 250. These are reported customer-specific terms, not published rates available to all buyers.
Rippling CEO Parker Conrad said his company’s distribution of coding work changed substantially: OpenAI’s share rose from 16 percent in May to 52 percent, while Anthropic’s fell from 57 percent to 18 percent. Conrad attributed the shift to cost. DataCamp CEO Jonathan Cornelissen said OpenAI’s offer could reduce his company’s AI costs by 80 to 90 percent, but that figure describes expected savings against its existing setup, not a 90 percent cut to OpenAI’s list price.
Prices published by the vendors also fell. Between June and September, OpenAI’s reported Sol model pricing moved from $5 input and $30 output per million tokens to $2 and $10; its Luna model went from $1 and $6 to $0.10 and $0.50. Anthropic’s Opus pricing moved from $5 and $25 to $4 and $20 per million tokens, while its Sonnet model launched at introductory pricing that the company later made permanent. Those comparisons involve different model releases, not simply unchanged models becoming cheaper.
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
Discounts Reshape AI Buying
The contest gives enterprise buyers more leverage as they negotiate contracts, compare providers and decide which tasks each model should handle. The reported Replit offer and the shifts described by Rippling show that competition can affect actual deployments, not only advertised prices. But one customer’s savings cannot be assumed to apply to another: workloads, negotiated rates, usage levels and contract commitments all differ.
List prices are only part of the bill. Token allowances, seat minimums, credits and usage limits can change the effective cost even when a monthly subscription price stays the same. The supplied source says Anthropic added capped, expiring API credits to Max and Team plans and requires 20 seats for Enterprise. It also says OpenAI reduced usage included with new $200 Pro subscriptions without changing the price. Buyers need to check those terms alongside model rates.
Products Worth Considering
Samsung 27-Inch Odyssey QD-OLED G8 4K 240Hz Gaming Monitor
Samsung 27 Odyssey OLED G5 (G50SF) QHD QD-OLED Gaming Monitor
Samsung 32-Inch Odyssey OLED G8 (G80SH) 4K Gaming Monitor
As an affiliate, we earn on qualifying purchases.
Months of Falling Model Prices
The pricing changes accumulated from June through September 2026. OpenAI and Anthropic introduced newer models at lower prices than some earlier models, while also changing prices for selected services and cached-token processing. Data scientist Max Woolf tracked the moves in an October 25 blog post; the source says its author checked the figures against vendor announcements and pricing pages retrieved October 26.
For companies using coding agents, cached-token rates may matter as much as headline input and output prices. Agents can repeatedly process prior conversation context. Woolf estimates cached tokens account for 96 to 99 percent of token spending in agentic coding, according to the supplied source. Anthropic has said its Fable 5.1 model costs around 25 percent less on typical workloads and up to around 45 percent less on highly agentic workloads; those are the company’s estimates, not independent guarantees.
Model-to-model comparisons also have limits. OpenAI described GPT-6.1 Sol as offering “near-Astra intelligence for a fifth of the price,” but buyers would need to test whether a newer model performs comparably on their own tasks. A lower price per token does not by itself establish lower total cost or equivalent results.
“OpenAI’s share of coding work rose from 16 percent in May to 52 percent, while Anthropic’s fell from 57 percent to 18 percent.”
— Parker Conrad, Rippling CEO
Deal Terms and Savings Vary
The public information does not establish the full terms of each reported contract, how long discounts last or whether they include minimum spending commitments. The Information article is paywalled, and the account supplied here also draws on public summaries and other reports. The reported 80-to-90-percent DataCamp saving is an expected total-cost comparison; the baseline details and calculation are not provided.
Replit’s relationship with Anthropic is also described differently by different accounts. Masad said the company was no longer working with Anthropic at an event, then told TechCrunch that Replit remained model agnostic and OpenAI was its preferred partner. A person familiar with the partnership told TechCrunch it was still in place. The exact scope of current use across providers is not clear from the available material.
It is also not clear whether lower-priced successor models deliver equivalent results for every customer, or how broadly the reported concessions are available. Contract terms and plan limits can change, so the cited prices and offers are a snapshot rather than a guarantee of current rates.
Buyers Compare Full Contract Costs
Companies weighing the offers will need to compare their own workloads across providers, including model performance, token use, cached-token rates, support and seat or credit rules. The reported deals suggest buyers may have grounds to revisit existing agreements, but the source material does not identify a common deadline or a general offer available to all customers.
Further clarity may come from vendor pricing pages and announcements, as well as customers disclosing the terms and duration of their contracts. For now, the measurable question is not only which model has the lower list price, but what each provider charges for the specific work a customer needs and how much usage its plan actually includes.
Key Questions
Is OpenAI offering every Anthropic customer a 90 percent discount?
No. The source material describes customer-specific offers and says DataCamp’s 80-to-90-percent figure is an expected reduction in total costs against its existing setup, not a discount from OpenAI’s published prices.
What customer changes were reported?
Rippling CEO Parker Conrad said OpenAI’s share of coding work at his company rose from 16 percent in May to 52 percent, while Anthropic’s share fell from 57 percent to 18 percent. Replit was also reported to have received a discount of more than 30 percent on a large contract.
Have OpenAI and Anthropic both cut model prices?
Yes. The supplied source compares vendor pricing from June to September 2026 and reports reductions across several model releases. Some comparisons are between successor models, so they do not show that the same model became cheaper while remaining unchanged.
What should businesses check beyond token prices?
They should review seat minimums, included usage, credit limits and expiry dates, cached-token pricing, contract duration and model performance on their own tasks. Those terms can affect total cost even when the headline monthly fee or token price looks lower.
Source: ThorstenMeyerAI.com
