The Price Battle For AI Customers: OpenAI Vs. Anthropic

🔍 Read the full analysis: The Price Battle For AI Customers: OpenAI Vs. Anthropic on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get smart everyday buys delivered free with Prime

• Fast, free delivery on millions of items
• Prime Video, Amazon Music and more included
• Member-only deals all year

Start your free Prime trial →Free trial for eligible customers · Cancel anytime

As an affiliate, we earn on qualifying purchases.Hide

OpenAI and Anthropic have cut prices on several AI models since June, and a report by The Information says OpenAI is also offering discounts to attract enterprise customers using Anthropic. The developments are changing costs for AI buyers, but a widely cited 80% to 90% savings estimate describes one company’s expected total cost reduction—not a discount of that size off OpenAI’s published prices.

Pricing pages and vendor announcements reviewed by ThorstenMeyerAI.com showed OpenAI’s Sol model moving 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. The smaller Luna model moved from $1 and $6 to $0.10 and $0.50. These figures compare different model releases, rather than a single unchanged model receiving a price cut.

Anthropic also lowered prices or changed costs across its model line. Its Opus model moved from $5 per million input tokens and $25 per million output tokens for Opus 5 to $4 and $20 for Opus 5.5. Anthropic said cache reads for Fable 5.1 were 75% cheaper; it also said that model could cost around 25% less on typical workloads and up to around 45% less on highly agentic workloads. Those workload savings are the company’s estimates, not independently verified results.

The Information reported October 24 that OpenAI is using discounts to pursue customers associated with Anthropic. A person cited in the reporting said Replit received an offer of more than 30% off a contract valued at over $100 million annually. DataCamp CEO Jonathan Cornelissen said he expected OpenAI’s pricing to lower his company’s costs by 80% to 90%. That figure refers to expected total savings against DataCamp’s existing setup, not a 90% reduction in OpenAI’s list price.

Lower Prices, Different Deal Terms

The competition matters because model prices affect the cost of using AI at scale, including in software development and other workloads that generate many tokens. A lower per-token price can change which models a company uses, but it does not by itself show that a new model performs as well as the one it replaces on a particular task.

Enterprise buyers also face a less visible question: what usage and contract terms come with a quoted price? The source reports that Anthropic added capped, expiring API credits to its Max and Team plans, while its Enterprise plan requires 20 seats. It also says OpenAI reduced the usage allowance included with new $200 Pro subscriptions without changing the monthly price. These changes mean buyers and individual users may see costs or access shift even when a headline price stays the same.

For companies, the reported offers give them grounds to review existing agreements and compare vendors on total cost and actual workload performance, not just public token rates. The reported customer changes suggest price can influence adoption, though they do not establish that price alone caused each shift.

How the Price Cuts Accumulated

The price changes described by ThorstenMeyerAI.com span June through September 2026. OpenAI reduced Luna pricing in July and cut Sol pricing in August before launching GPT-6 Sol and Luna at lower rates in September. Anthropic introduced Sonnet 5 at an introductory price in June and made that price permanent in August; it launched Opus 5.5 in September. The source says it checked the listed figures against vendor announcements and pricing pages on October 26.

Some cost changes involve cached tokens, rather than the main input and output rates. In agentic coding, a system may reuse conversation context across repeated steps. Data scientist Max Woolf estimated that cached tokens account for 96% to 99% of token spending in agentic coding. That is his estimate, and the share will depend on the workload. It helps explain why cache pricing can matter even when a model’s headline rates do not change.

Reported customer movement offers another measure of the competition. 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%. Replit CEO Amjad Masad said at an October 16 event that his 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 told TechCrunch it was still in place.

“I think ChatGPT is better now.”

— Amjad Masad, Replit CEO, speaking at TechCrunch Disrupt on October 16

Deal Details Still Matter

The full terms of the reported enterprise offers are not public, and the available information does not establish how many customers received comparable discounts. The Information’s article is paywalled; the source material says its account was checked against a public summary and two secondary reports. The reported Replit discount, DataCamp savings estimate and bank seat arrangement should not be treated as a single standard offer.

It is also unclear whether the customer shifts will persist or how much they reflect price, model performance, product features or other contract terms. Replit’s public comments and the account from a person familiar with its partnership do not fully resolve the status or scope of its relationship with Anthropic.

Finally, comparisons between successive models do not establish equivalent performance. The published token prices are verifiable figures, but real costs depend on usage, caching, plan allowances and task results. The source does not provide enough detail to calculate a consistent, independently verified comparison of total costs across all customers.

Watch Renewals and Usage Limits

Companies considering a change can compare current contracts with vendor offers and test competing models on representative workloads before shifting substantial work. That comparison should account for input and output tokens, cache rates, usage caps, credits, seat minimums and renewal terms, rather than relying on a headline discount alone.

The next evidence will come from more disclosed contract terms, customer usage data and vendor pricing updates. Buyers should also check plan allowances when renewing or starting a subscription: the monthly fee may remain the same even as included usage or credits change. Whether the recent cuts and offers lead to durable customer shifts remains to be seen.

Key Questions

Did OpenAI cut its prices by 80% to 90%?

No. That range was DataCamp CEO Jonathan Cornelissen’s estimate of expected total savings against his company’s current setup. It was not reported as a reduction of that size from OpenAI’s list price.

How much did OpenAI’s listed model prices change?

According to pricing figures 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 $0.10 and $0.50. These are prices for successive model versions.

Did Anthropic also reduce prices?

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. Anthropic also reduced cache-read prices for Fable 5.1 and published estimates of lower costs on certain workloads.

Are companies switching from Anthropic to OpenAI?

Some reported usage has shifted. Rippling CEO Parker Conrad said OpenAI’s share of coding work at his company increased from 16% in May to 52%, while Anthropic’s fell from 57% to 18%. Replit’s partnership status is less clear: public comments called OpenAI its preferred partner, while a source told TechCrunch the partnership with Anthropic remained in place.

What should AI buyers compare beyond token prices?

Buyers should compare model performance on their own tasks, cache pricing, included usage, credits and expiration rules, seat minimums, and contract terms. A lower list price does not necessarily mean a lower total bill.

Source: ThorstenMeyerAI.com