🔍 Read the full analysis: A Buyer’s Guide To OpenAI And Anthropic’s AI Price Offers on ThorstenMeyerAI.com
Get business pricing on tech for your team
- 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 models, while reporting describes OpenAI offering large customer-specific discounts to win business from Anthropic. The reported 80% to 90% savings refer to one buyer’s expected overall cost reduction, not a general discount off OpenAI’s list prices. Buyers should compare workload-specific costs, contract terms, usage limits and model performance.
OpenAI and Anthropic have cut prices for several AI models in recent months, and a report by The Information says OpenAI is also offering customer-specific discounts to win business from Anthropic. The changes give companies more room to negotiate, but quoted savings vary by deal and do not establish a universal discount or a like-for-like reduction in total costs.
On October 26, ThorstenMeyerAI.com said it checked a price timeline against vendor announcements and current pricing pages. In that account, OpenAI’s GPT-6.1 Sol was listed at $2 per million input tokens and $10 per million output tokens, compared with $5 and $30 for an earlier Sol model in June. Its Luna line moved from $1 and $6 to $0.10 and $0.50. These comparisons cover different model versions, rather than showing the same model’s price falling while its capabilities remain unchanged.
Anthropic also lowered prices across model releases, according to the same account. Opus pricing moved from $5 input and $25 output per million tokens for Opus 5 to $4 and $20 for Opus 5.5. Anthropic also cut cache-read prices for some models. The company’s claims that caching reduces typical workload costs by particular percentages are vendor estimates; actual costs depend on how a customer’s software uses the models.
The Information reported on October 24 that OpenAI was using discounts to attract customers from Anthropic. The account summarized by ThorstenMeyerAI.com includes a reported discount of more than 30% on a Replit contract worth over $100 million a year, and a US bank that received 500 ChatGPT seats for less than $7 per seat. Replit CEO Amjad Masad said at TechCrunch Disrupt on October 16 that the company was no longer working with Anthropic and that he thought ChatGPT was better. He later told TechCrunch that Replit remained model-agnostic and OpenAI was its “preferred partner”; a person familiar with the partnership said it remained in place.
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
What Buyers Should Compare
Lower list prices can reset negotiations, especially for organizations buying large volumes of API access or enterprise seats. But a price per million tokens or per seat is only one part of the bill. Usage limits, cached-token rates, contract length, minimum seat counts and credits can change the effective cost substantially.
The reported savings also describe different things. DataCamp CEO Jonathan Cornelissen reportedly said OpenAI’s offer could reduce his company’s costs by 80% to 90%. That is an expected saving against the company’s existing setup, not a general 80% to 90% reduction from OpenAI’s published list price. The figure is a buyer’s estimate reported by others, and the source material does not provide the underlying contract or calculation.
Companies should test models on their own tasks before switching. A lower-priced successor may not perform equally well on a particular coding, support or research workflow. Where practical, routing work across more than one provider can give buyers options, though it also adds integration and evaluation work. Individual subscribers face a related issue: a monthly fee can stay the same while included usage changes.
Products Worth Considering
Samsung 27 Odyssey OLED G5 (G50SF) QHD QD-OLED Gaming Monitor
Bose SoundLink Flex Bluetooth Speaker (2nd Gen), Portable Waterproof Speaker, Black
Sony WH-1000XM4C Wireless Noise-Cancelling Headphones, Black
As an affiliate, we earn on qualifying purchases.
How Model Prices Shifted
The price changes followed a series of model launches and adjustments from June through September 2026. The timeline cited by ThorstenMeyerAI.com includes OpenAI price reductions, Anthropic setting an introductory Sonnet price as permanent, and new model releases from both companies. Because many comparisons pair an older model with a newer one, the headline price alone does not establish that buyers receive identical capability for less.
Cached tokens matter for some agent workloads. Software agents may repeatedly process prior conversation or task context, so cache pricing can affect the total bill even when standard input and output rates stay unchanged. Anthropic said its Fable 5.1 could cost around 25% less on typical workloads and up to around 45% less on highly agentic workloads. Those are Anthropic’s estimates, not independently verified savings for every customer.
Plan terms have shifted alongside API rates. The source account says Anthropic added capped, expiring API credits to Max and Team plans and that its Enterprise plan has a 20-seat minimum. It also says OpenAI reduced usage included in new $200 Pro subscriptions without changing the monthly price. These details mean buyers need to compare the terms applying to their own plan and contract, not just prices shown in a headline or older offer.
“I think ChatGPT is better now.”
— Amjad Masad, Replit CEO, at TechCrunch Disrupt
Deal Terms and Savings Still Vary
Not all reported offers are public or independently verifiable from the material provided. The Information’s article is paywalled; the account here relies on its public summary and secondary coverage. Full contract terms, eligibility, duration and usage conditions for the reported discounts are not included.
The Replit accounts also require care: Masad said the company was no longer working with Anthropic, then told TechCrunch it remained model-agnostic, while a source familiar with the partnership said it was still in place. The available material does not spell out the current contract scope or whether the arrangement is exclusive. It also does not provide the bank’s full seat agreement or enough detail to calculate its total cost.
Published rates and plan allowances can change. Comparisons between model generations do not by themselves show equivalent quality, and the reported customer savings cannot be generalized to other buyers. The material also does not establish how long each offer will remain available.
Check Offers Against Real Workloads
For a purchase or renewal, buyers can request a written quote and confirm the applicable model versions, token rates, cached-token pricing, seat minimums, included usage, credit expiry and overage charges. They should run representative tasks across providers and compare both output quality and total cost over a defined period.
Watch for changes to vendor pricing pages and plan terms, as well as new reporting on customer contracts and model adoption. Until further details emerge, the clearest conclusion is that buyers have more pricing leverage than list prices alone suggest—but the size of any saving depends on the offer, the workload and the fine print.
Key Questions
Does OpenAI offer every customer an 80% to 90% discount?
No. The reported 80% to 90% figure is DataCamp CEO Jonathan Cornelissen’s estimate of expected savings against his company’s current setup. It is not a general discount from OpenAI’s list prices.
How much did the cited model prices change?
According to the pricing account dated October 26, 2026, OpenAI’s Sol line moved from an earlier model’s $5 input and $30 output per million tokens to GPT-6.1 Sol at $2 and $10. Luna moved from $1 and $6 to $0.10 and $0.50. These are comparisons between different model versions.
Did Replit leave Anthropic for OpenAI?
Replit CEO Amjad Masad said the company was no longer working with Anthropic and later called OpenAI its preferred partner, while also saying Replit remained model-agnostic. A person familiar with the partnership told TechCrunch it was still in place. The precise relationship and contract scope remain unclear.
What should businesses check before switching providers?
Compare performance on representative tasks, total usage-based costs, cache rates, usage limits, seat minimums, credits, contract duration and overage charges. A lower headline rate does not guarantee a lower bill or equivalent results.
Source: ThorstenMeyerAI.com
