🔍 Read the full analysis: How OpenAI And Anthropic Are Competing For AI Customers On Price 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 lowered prices on several models, while reporting describes OpenAI offering discounts to win customers from Anthropic. A reported 80% to 90% cost saving refers to one buyer’s expected total savings, not a general discount off OpenAI’s list prices. Customers also need to track usage limits, credits and seat requirements, which can change the effective cost.
OpenAI and Anthropic have cut prices on several AI models since June, while OpenAI is also offering reported customer-specific discounts as it competes with Anthropic for business. The changes could lower bills for companies buying large amounts of AI, but list prices do not show the full cost: usage allowances, credits and contract terms are changing too.
According to pricing information checked by ThorstenMeyerAI.com on October 26 against vendor announcements and pricing pages, OpenAI’s Sol model line 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. Its smaller Luna line moved from $1 and $6 to 10 cents and 50 cents per million tokens. These figures refer to different model releases, not necessarily a price cut for an unchanged model.
Anthropic also reduced prices across releases. Its Claude Opus model moved from $5 input and $25 output per million tokens for Opus 5 in July to $4 and $20 for Opus 5.5. Anthropic also cut cache-read prices on some models. Cached tokens can make up much of the spending in agentic coding workflows, according to data scientist Max Woolf, who tracks model pricing. Anthropic’s estimates that particular releases cost less on typical or agentic workloads are vendor claims, and results depend on how customers use the models.
Separate reporting by The Information, published October 24, described OpenAI offering discounts to customers that had used Anthropic. The reported deals include a discount of more than 30% for Replit on a contract worth more than $100 million a year, according to a person cited in the report. DataCamp CEO Jonathan Cornelissen said he expected OpenAI’s offer to reduce his company’s costs by 80% to 90%. That figure describes expected savings against DataCamp’s existing setup, not a general discount from OpenAI’s list 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
For companies using AI across software development, customer support or other business tasks, lower token prices can reduce a recurring operating expense. Customer-specific offers add another lever: buyers may be able to negotiate based on workload, volume and the alternative providers they can use. The reported deals also show that competition is not limited to model benchmarks; providers are competing for the contracts and usage that determine where business customers build their workflows.
Still, a lower price per token does not automatically mean a lower total bill. A new model may perform differently on a company’s tasks, and the amount of input, output and cached data can vary by workflow. Businesses need to compare effective cost and performance on their own workloads, including any usage caps and contract conditions, rather than relying on a single headline price or a reported savings percentage.
Products Worth Considering
Samsung 27-Inch Odyssey QD-OLED G8 4K 240Hz Gaming Monitor
Samsung 32-Inch Odyssey OLED G8 (G80SH) 4K Gaming Monitor
Samsung 27 Odyssey OLED G5 (G50SF) QHD QD-OLED Gaming Monitor
As an affiliate, we earn on qualifying purchases.
How the Price Cuts Accumulated
The reductions came in a series of launches and pricing changes from June to September 2026. Anthropic introduced Sonnet 5 at an introductory price in June and made that pricing permanent in August, according to the source material. OpenAI lowered Luna pricing in July and cut Sol pricing in August before introducing GPT-6 Sol and Luna at lower prices in September. Anthropic also released Opus 5.5 in September at a lower listed price than Opus 5.
These comparisons require care: many price points belong to successor models, rather than the same model becoming cheaper. Whether a successor offers comparable quality for a particular task cannot be settled by the price table. The source also says OpenAI reduced the usage included in new $200 Pro subscriptions without changing the monthly fee, while Anthropic added capped, expiring API credits to some plans. Those terms can affect what customers receive for the advertised price.
Contract details matter as well. The source says Anthropic’s Enterprise plan has a 20-seat minimum, which may put it out of reach for some smaller firms. It also reports that a U.S. bank with more than $7 billion in assets had its ChatGPT seats doubled to 500 at no extra charge. These examples illustrate why seat counts, eligibility and usage rules belong in any comparison of competing offers.
“I think ChatGPT is better now.”
— Replit CEO Amjad Masad, speaking at TechCrunch Disrupt on October 16
What the Reported Savings Cover
The full terms of several customer offers are not public. The Information’s article is paywalled, and the available source material provides only selected details from its reporting and public summaries. It is not clear how broadly OpenAI is making these discounts available, how long they last, or what usage commitments and other conditions apply.
The 80% to 90% figure is one customer’s expected total savings, not a verified result across buyers. The reported Replit discount and the bank’s expanded seats also do not establish a standard offer for other customers. Separately, model-to-model price comparisons do not confirm equivalent performance, and vendor estimates of workload savings have not been independently established in the supplied material.
Some subscription and credit changes are reported by the source, but the scope, timing and terms may differ by plan or customer. Buyers should verify current pricing and contract details directly with the vendors before estimating savings.
What Buyers Should Track
Companies considering a change can compare vendors using representative tasks and their actual usage data, then ask providers for written terms covering rates, included usage, credit expiry, minimum seats and renewal pricing. Testing more than one provider can also help buyers avoid relying on a single vendor’s price or model-performance claims.
The next useful evidence will be whether reported discounts become available to a broader set of customers and how the providers’ prices and plan limits change. For users on subscriptions, the monthly fee alone may not reveal a change in value: check included usage and caps as well as the advertised price. The available reporting does not establish the final terms or duration of the customer offers described.
Key Questions
Did OpenAI cut its prices by 90%?
No general 90% list-price cut is established here. The figure refers to DataCamp CEO Jonathan Cornelissen’s estimate of total savings from OpenAI’s offer compared with his company’s existing setup.
How much did OpenAI’s Sol and Luna prices change?
According to vendor pricing information checked by ThorstenMeyerAI.com, Sol moved 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 10 cents and 50 cents. These are comparisons across model releases.
What discounts did OpenAI reportedly offer customers?
The Information reported a discount of more than 30% for Replit on a contract worth over $100 million a year. The source also describes other customer-specific terms, but the full offer details are not public.
Why can the list price differ from a company’s actual AI bill?
Actual spending depends on workload and token use, including cached tokens. It can also depend on subscription allowances, credit limits and expiry, seat minimums, and negotiated contract terms.
Source: ThorstenMeyerAI.com
