🔍 Read the full analysis: A Practical Look At OpenAI And Anthropic’s AI Pricing Battle on ThorstenMeyerAI.com
Get hardware and tech essentials 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 prices for several AI models since June, while OpenAI is also offering large customer-specific discounts, according to reporting cited by the source. Customer usage appears to be shifting, but the widely repeated 80% to 90% savings figure describes one buyer’s expected total cost reduction, not a general OpenAI price cut. Contract terms, included usage and model performance remain important parts of the comparison.
OpenAI and Anthropic have lowered prices on several AI models since June, and new reporting says OpenAI is using customer-specific discounts to compete for Anthropic accounts. The changes are affecting both per-token rates and subscription terms, giving businesses more room to negotiate while making headline prices an incomplete guide to total cost.
Based on vendor pricing pages and announcements reviewed by the source on October 26, OpenAI’s Sol model pricing 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 model moved from $1 and $6 to $0.10 and $0.50 per million tokens. These figures compare different model releases; they do not show that an unchanged model simply became cheaper.
Anthropic also reduced prices. 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. The source reports that cache reads for Opus 5.5 are 60% cheaper. Anthropic has also said Opus 5.5 costs 40% less on typical workloads than Opus 5, and that its Fable 5.1 model can cost less on agentic workloads. Those workload savings are Anthropic’s estimates, not independent guarantees.
The Information reported on October 24 that OpenAI has offered discounts to win business from Anthropic. The source cites a reported discount of more than 30% for Replit on a contract valued at over $100 million a year, as well as a U.S. bank receiving ChatGPT seats for less than $7 each. The details and terms of those arrangements have not been independently established in the source material.
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 Rates Reshape Enterprise AI Deals
For companies paying for large volumes of model use, lower token rates can reduce operating costs, but the value depends on how much work a model handles, how often it is used and whether it performs well on a company’s tasks. Customer-specific discounts add another variable: two buyers may pay different amounts for access to the same service.
Reported shifts at customers suggest pricing is influencing vendor choices. 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 company’s account, not evidence of a market-wide shift. Still, it illustrates how quickly companies may change which models they use when costs or performance change.
Buyers also need to compare more than price per token. Seat minimums, usage limits, credits and expiration rules can change the effective cost of a contract. A lower list price may not produce savings if a plan includes less usage, requires more seats than a company needs or charges separately for high-volume work.
Products Worth Considering
LG UltraGear 32GX870A-B 32-inch 4K OLED Gaming Monitor
LG 24G414B-B 24-Inch UltraGear Full HD IPS Gaming Monitor
LG UltraGear 27G640A-B 27-Inch QHD IPS Gaming Monitor, 300Hz, 1ms, G-SYNC Compatible, FreeSync Premium, DisplayHDR 400
As an affiliate, we earn on qualifying purchases.
A Four-Month Run of Price Changes
The reductions followed a sequence of model releases and pricing adjustments between June and September 2026. The source compared pricing announcements and current vendor pages with figures published by data scientist Max Woolf, who tracks large language model prices. The sequence included Anthropic making Sonnet 5’s introductory price permanent, OpenAI reducing Luna’s rate, and later releases from both companies at lower listed prices.
Some changes involved new models replacing earlier versions, rather than a price cut for the same product. That distinction matters because a lower rate does not by itself establish equal or improved performance. OpenAI described GPT-6.1 Sol as offering “near-Astra intelligence for a fifth of the price,” but buyers would need to test the model on their own workloads to judge whether it meets their needs.
Pricing for cached tokens is another factor in some applications. Coding agents can repeatedly process earlier conversation context, and the source cites Woolf’s estimate that cached tokens account for 96% to 99% of token spending in agentic coding. That estimate helps explain why cache pricing can affect bills even when standard input and output rates stay unchanged.
Actual Savings Depend on Contract Terms
The reported 80% to 90% savings should not be read as a general discount available to OpenAI customers. It is DataCamp CEO Jonathan Cornelissen’s estimate of savings against his company’s current setup. The source describes a separate reported discount of more than 30% for Replit, but does not provide the full contract terms.
It is also unclear how widely the reported offers are available, how long they last, or what usage commitments, service conditions and renewal prices they include. 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.
Plan limits are changing alongside rates. The source says Anthropic added capped, expiring API credits to Max and Team plans, and that its Enterprise plan requires 20 seats. It also reports that OpenAI reduced included usage for new $200 Pro subscriptions without changing the monthly price. The precise impact will vary by plan and customer, and the source does not provide a complete comparison of all current contract terms.
Buyers Will Test Offers and Limits
Companies considering a change will need to compare written proposals against their actual usage, including token rates, cache charges, seat counts, usage allowances, credit expiration and renewal terms. Testing candidate models on representative work can help buyers distinguish a lower price from a lower total cost that still meets performance requirements.
Further vendor announcements, updated pricing pages and customer disclosures may clarify how broadly the reported discounts apply. For now, the available material points to continued price competition, but it does not establish a single market-wide rate or a universal level of savings. Buyers should treat quoted savings as specific to the customer and contract 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 that OpenAI’s offer could reduce his company’s total costs by 80% to 90% compared with its current setup.
How much did OpenAI’s listed model prices change?
The source says Sol pricing 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 $0.10 and $0.50. These comparisons involve different model releases.
Are Anthropic’s prices falling too?
Yes. The source reports Opus pricing moving from $5 input and $25 output per million tokens for Opus 5 to $4 and $20 for Opus 5.5, alongside a lower cache-read price. Anthropic’s workload cost-reduction figures are company estimates.
What should businesses compare beyond token prices?
They should review total expected usage costs, seat minimums, included allowances, credit caps and expiration dates, cache rates, contract duration and renewal terms. Model performance on the company’s own tasks also matters.
Source: ThorstenMeyerAI.com
