🔍 Read the full analysis: The OpenAI And Anthropic Price War: Deals, Details, And Decisions on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
OpenAI and Anthropic have cut several model prices since June, while reporting describes OpenAI offering discounts to win customers from Anthropic. The widely cited 80% to 90% savings figure refers to one executive’s expected total cost reduction, not a general discount off OpenAI’s list prices. Contract terms and usage limits are changing too, so buyers need to compare their actual workloads and total costs.
OpenAI and Anthropic have cut prices on several AI models since June, and OpenAI is offering some customers discounts as it competes for business associated with Anthropic, according to an October 24 report by The Information. The changes matter to companies buying AI services, but the reported 80% to 90% savings refers to one customer’s expected overall cost reduction, not a general discount on OpenAI’s published prices.
Pricing pages and vendor announcements reviewed by ThorstenMeyerAI.com show OpenAI’s GPT-6.1 Sol priced at $2 per million input tokens and $10 per million output tokens in September, down from $5 and $30 for the earlier Sol model in June. Its smaller Luna model fell from $1 and $6 to $0.10 and $0.50. The comparison is between model generations, rather than a like-for-like reduction for an unchanged model.
Anthropic also lowered prices over the period. Its Opus line moved from $5 per million input tokens and $25 per million output tokens for Opus 5 in July to $4 and $20 for Opus 5.5. Anthropic also cut prices for reading cached input tokens on some models. The company says those cache changes reduce costs on typical and highly agentic workloads, but those estimates are vendor claims and may not match every customer’s usage.
The Information’s reporting describes several customer offers and shifts in model use. Replit reportedly received a discount of more than 30% on a contract worth over $100 million a year. A U.S. bank with more than $7 billion in assets was offered 500 ChatGPT seats at no extra charge, according to the source article. Rippling CEO Parker Conrad said OpenAI’s share of his company’s coding work rose from 16% in May to 52%, while Anthropic’s share fell from 57% to 18%. Those figures describe Rippling, not the broader market.
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 Bills Depend on Usage
Lower per-token prices can reduce spending for businesses that process large volumes of AI requests, especially when coding agents repeatedly send long conversation histories. But list prices are only part of the bill: model choice, cached-token charges, negotiated discounts, included usage and contract rules all affect what a customer pays.
The reported deals also show that competition is happening at the contract level, not only through public price lists. A discount tied to a large annual commitment may not be available to smaller buyers. Companies that can test multiple providers may gain negotiating leverage, while organizations locked into one platform could have fewer options. For individual subscribers, a monthly fee can stay the same even as included usage changes.
That makes the headline savings figures difficult to compare without a baseline. DataCamp CEO Jonathan Cornelissen’s reported estimate of 80% to 90% lower costs concerns his expected total spending under OpenAI’s offer compared with his current setup. It is not evidence that OpenAI cut its posted prices by that amount, nor a guaranteed saving for other customers.
Products Worth Considering
LG UltraGear 32GX850A-B 32-Inch 4K UHD Glossy OLED Gaming Monitor
Samsung 27-inch Odyssey OLED G6 QD-OLED QHD Gaming Monitor
Samsung 27 Odyssey OLED G5 (G50SF) QHD QD-OLED Gaming Monitor
As an affiliate, we earn on qualifying purchases.
A Fast-Moving Price Timeline
The reductions followed a sequence of launches and pricing changes from June through September 2026. Anthropic introduced Sonnet 5 at an introductory rate in June and later made that price permanent, according to the source material. OpenAI cut Luna pricing in July and reduced Sol pricing temporarily in August, before launching GPT-6 models at lower listed prices in September. Anthropic also released Opus 5.5 and made changes to cache pricing.
Data scientist Max Woolf published a timeline of the cuts on October 25. ThorstenMeyerAI.com says it checked the figures against vendor announcements and pricing pages retrieved on October 26. The price comparisons largely involve newer models replacing older models, so buyers cannot assume the same task will produce the same quality or cost simply by comparing two price rows.
Some of the less visible changes involve product terms. The source reports that Anthropic added capped, expiring API credits to its Max and Team plans and requires 20 seats for Enterprise. It also says OpenAI reduced the usage included with new $200 Pro subscriptions without changing the subscription price. Such limits can affect a buyer’s effective cost even when the advertised price is unchanged.
“OpenAI’s share of coding work rose from 16% in May to 52%, while Anthropic’s fell from 57% to 18%.”
— Parker Conrad, Rippling CEO, as reported by The Information
Deal Terms Still Vary
The public information does not establish the full terms of the reported customer offers, including how long discounts last, what usage commitments apply or which services are included. The Information article is paywalled, and some details in the source account come through public summaries and follow-up reporting.
It is also unclear how representative the reported deals are. Replit’s discount, the bank’s seat offer and DataCamp’s projected savings involve distinct customers and arrangements. Replit CEO Amjad Masad said the company was model agnostic and called OpenAI its preferred partner; a person familiar with the partnership told TechCrunch it remained in place. These statements do not by themselves clarify the contract’s precise terms or Replit’s future model mix.
Finally, the comparison of prices across model generations does not show whether newer models perform as well or better on every customer task. Published rates and vendor cost estimates cannot resolve that question without workload-specific testing.
Compare Workloads and Contracts
Companies considering a renewal or new AI contract can use the price changes as a reason to request updated offers, but should compare total cost on representative workloads rather than rely on a per-token rate or a headline savings percentage. That comparison should account for model quality, cached-token use, seat counts, usage allowances, credit expiry and any minimum commitment.
Further details about the reported offers may clarify how discounts are structured and whether they are temporary or tied to specific volumes. Customers will also need to track changes to subscription limits and test model performance as providers update their product lines. The terms and wider effects of the reported customer shifts remain developing.
Key Questions
Did OpenAI cut its prices by 90%?
No general 90% list-price cut is established by the source material. The figure refers to DataCamp CEO Jonathan Cornelissen’s expected overall cost savings from an OpenAI offer compared with his company’s current setup.
How much did the listed model prices change?
The source reports OpenAI Sol moving from $5 input and $30 output per million 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. These are comparisons across model generations.
Are companies switching from Anthropic to OpenAI?
The reporting describes shifts at individual businesses, including Rippling and Replit. It does not establish a broad market-wide migration. Replit’s CEO said OpenAI was its preferred partner while also saying the company remained model agnostic.
What should AI buyers compare besides list price?
Buyers should compare charges for input, output and cached tokens, as well as discounts, seat minimums, usage allowances, credit expiration, contract duration and model performance on their own tasks.
Source: ThorstenMeyerAI.com
