🔍 Read the full analysis: How The OpenAI-Anthropic Price War Is Changing AI Subscriptions 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 repeatedly cut API prices, while OpenAI is also reportedly offering large customer-specific discounts to win business from Anthropic. Reported savings vary by customer and should not be confused with across-the-board list-price cuts; subscription allowances and contract terms are changing alongside prices.
OpenAI and Anthropic have cut model prices several times since June, and OpenAI is reportedly offering additional discounts to win customers from Anthropic, according to an October 24 report by The Information. The pricing shifts are changing enterprise deals and the cost of using AI, but buyers also face changes in subscription usage allowances, credits and seat requirements.
Vendor pricing pages reviewed by ThorstenMeyerAI.com on October 26 show that OpenAI’s Sol model line moved from $5 input and $30 output per million tokens in June to $2 and $10 for GPT-6.1 Sol in September. Its Luna line moved from $1 and $6 to $0.10 and $0.50 per million tokens. These are prices for different model versions, not simply reductions for an unchanged model.
Anthropic also reduced prices across its model line. Opus pricing went from $5 input and $25 output per million tokens for Opus 5 in July to $4 and $20 for Opus 5.5. Anthropic also reduced cache-read prices for some models. The company says those changes can lower costs further for workloads that repeatedly reuse prompts; its estimates of savings for typical and agentic workloads are vendor claims, not independent measurements.
The Information’s reporting describes customer-specific offers, not a general 90% cut to OpenAI’s public prices. Replit reportedly received a discount of more than 30% on a contract worth over $100 million a year. DataCamp CEO Jonathan Cornelissen said the offer could reduce his company’s overall AI costs by 80% to 90% against its existing setup. Those figures refer to expected total savings for particular customers and should not be read as a standard discount available to every buyer.
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
Why AI Buyers Have More Leverage
For companies buying AI at scale, falling public prices and competition for major accounts can strengthen their position in contract negotiations. A reported discount of more than 30% on a large contract illustrates how negotiated terms may differ from a vendor’s published rates. However, the reported customer examples do not establish that similar offers are available to all organizations.
The headline token price is only one part of the bill. Cached input, included usage, seat minimums and credit expiration rules can change the effective cost of a product even when its monthly subscription fee stays the same. Teams should compare spending against their own workloads and contract terms, rather than assuming that a lower list price means a proportionate reduction in total costs.
Competition can also affect which systems companies use. 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%. He attributed the shift to cost. That is one company’s reported experience, not evidence that customers generally are moving in the same direction.
Products Worth Considering
LG UltraGear 32GX850A-B 32-Inch 4K UHD Glossy OLED Gaming Monitor
Samsung 27-Inch Odyssey QD-OLED G8 4K 240Hz Gaming Monitor
Samsung 32-Inch Odyssey OLED G8 (G80SH) 4K Gaming Monitor
As an affiliate, we earn on qualifying purchases.
From Token Cuts to Contract Offers
The price changes accumulated over several months. OpenAI reduced Luna pricing in July and cut Sol pricing in August before introducing lower prices for GPT-6 Sol and Luna in September. Anthropic set introductory Sonnet 5 pricing as permanent in August, then introduced Opus 5.5 at lower listed rates than Opus 5. The figures cited here come from vendor pricing pages and announcements reviewed on October 26 by ThorstenMeyerAI.com.
Model-version changes make comparisons less direct: a cheaper successor may not perform identically on a buyer’s tasks. OpenAI described GPT-6.1 Sol as offering “near-Astra intelligence for a fifth of the price.” That is the company’s characterization; customers need to test models on their own work to judge whether the capability and cost trade-off suits them.
Subscription terms have shifted too. The source report says Anthropic added capped, expiring API credits to Max and Team plans and set a 20-seat minimum for Enterprise. It also says OpenAI reduced usage included with new $200 Pro subscriptions without changing the monthly price. These details mean that the effective value of a plan depends on how much and what kind of usage it includes, not just the fee.
“OpenAI’s share of the company’s coding work rose from 16% in May to 52%, while Anthropic’s fell from 57% to 18%.”
— Parker Conrad, Rippling CEO
What the Savings Figures Cover
The reported 80% to 90% saving for DataCamp is an estimate of total costs compared with its existing setup. The publicly available account does not establish the offer’s complete terms or how those projected savings were calculated. Likewise, the reported discount for Replit is tied to a specific contract and does not show what other customers could negotiate.
It is also unclear how widely the reported offers are being made, how long they last, or what minimum commitments and usage conditions apply. Replit’s relationship with OpenAI was described differently by public comments and sources cited by TechCrunch, though those accounts indicate a partnership remained in place. The cited pricing comparisons also involve successive model versions, so they do not by themselves show whether a cheaper model produces equivalent results for a particular task.
What Buyers Should Track Next
Companies considering a renewal or new AI contract can compare vendor quotes against actual usage, including input and output tokens, cached-token rates, included allowances, credits and seat minimums. Testing candidate models on representative tasks can help distinguish a lower price from a lower total cost for the same work. Where practical, buyers can also examine whether workloads can be routed across more than one provider.
Users and administrators should check subscription limits as well as monthly fees. The source material does not identify a common deadline for further price changes or specify when the reported enterprise offers expire. Buyers will need to follow vendor pricing pages and contract notices for updated terms, while watching whether public cuts continue and whether negotiated discounts become more widely available.
Key Questions
Are OpenAI’s reported 80% to 90% discounts available to everyone?
No such general discount is established. The 80% to 90% figure refers to DataCamp CEO Jonathan Cornelissen’s estimate of total savings against the company’s current setup. It is not a published reduction in OpenAI’s standard prices.
How much did OpenAI’s listed Sol and Luna prices change?
According to pricing pages reviewed on October 26, 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 $0.10 and $0.50. The comparisons involve different model versions.
Did Anthropic cut its prices too?
Yes. The cited figures put Opus at $5 input and $25 output per million tokens for Opus 5 in July, compared with $4 and $20 for Opus 5.5. Anthropic also changed cache-read pricing for some models.
Why should subscribers check usage limits if the monthly fee has not changed?
A plan’s value can change if its included usage, credits or seat rules change. The source report says OpenAI reduced usage included with new $200 Pro subscriptions without changing the fee, so customers should review current plan terms rather than relying only on the monthly price.
Source: ThorstenMeyerAI.com
