OpenAI And Anthropic Face Off Over AI Customers And Pricing

🔍 Read the full analysis: OpenAI And Anthropic Face Off Over AI Customers And Pricing on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get privacy and security gear 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 is offering customer-specific discounts to win business from Anthropic, while both companies have cut prices for several AI models, according to reporting summarized by ThorstenMeyerAI.com on October 26. The deals could reduce costs for some enterprise buyers, but the reported 80% to 90% savings describe one company’s expected total bill reduction, not a standard discount available to all customers.

The Information reported on October 24 that OpenAI had offered discounts to customers using Anthropic models. The published account, as described by ThorstenMeyerAI.com, includes a reported discount of more than 30% for Replit on a contract worth over $100 million annually, and a deal giving a U.S. bank with more than $7 billion in assets 500 ChatGPT seats at no additional charge. The source does not provide the full contracts or explain all conditions attached to either arrangement.

Customer accounts indicate that some companies are shifting workloads. 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%. Conrad attributed the change to cost. DataCamp CEO Jonathan Cornelissen said OpenAI’s offer was expected to reduce his company’s costs by 80% to 90%. That figure compares the expected total bill with DataCamp’s existing setup; it is not a 90% reduction from OpenAI’s public list prices.

Public prices have also moved. OpenAI’s GPT-6.1 Sol is listed at $2 per million input tokens and $10 per million output tokens, compared with $5 and $30 for an earlier Sol offering in June. Its Luna model went from $1 and $6 to $0.10 and $0.50. Anthropic’s Opus pricing moved from $5 and $25 per million input and output tokens for Opus 5 to $4 and $20 for Opus 5.5. These are comparisons across different model releases, not necessarily price cuts for an unchanged product.

Lower Bills, More Complex Deals

For companies buying AI services, the combination of public price reductions and private offers may strengthen their ability to negotiate. Buyers can compare vendors, request competing bids and test whether a lower-priced model performs well on their own tasks. The reported shift at Rippling suggests that price can influence which models employees use, but one company’s experience does not establish a broader market-wide migration.

List prices alone do not show the full cost. As the source notes, cached tokens and usage patterns can affect bills for coding agents, which repeatedly process conversation context. Anthropic says its Fable 5.1 model can cost about 25% less on typical workloads and up to about 45% less on highly agentic workloads; those are the company’s estimates, not independently verified savings for every customer. Buyers need workload-specific tests and contract terms to compare actual costs.

Plan limits also matter to individual users and smaller organizations. Anthropic added capped, expiring API credits to some Max and Team plans, while its Enterprise plan has a 20-seat minimum, according to the source. OpenAI reduced usage included with new $200 Pro subscriptions without changing the subscription price. A lower token rate or unchanged monthly fee can therefore coincide with a change in how much service a customer receives.

A Summer of Model Price Cuts

The reported competition follows a series of launches and pricing revisions from June through September 2026. OpenAI and Anthropic introduced successor models at lower listed rates than some earlier models, and both changed prices or allowances in the months that followed. Data scientist Max Woolf published a timeline of the price changes on October 25. ThorstenMeyerAI.com says it checked the listed figures against vendor announcements and pricing pages retrieved the next day.

Comparisons require care: the timeline covers different model generations, introductory prices and changes to cached-token rates. Anthropic made Sonnet 5’s introductory price permanent, while OpenAI announced temporary pricing for one model before later releases. Anthropic also cut cache-read prices for Fable 5.1. The source attributes claims about savings on typical and agentic workloads to Anthropic itself, and says buyers must test model quality against their own needs rather than assume a cheaper successor is equivalent.

Replit’s relationship with OpenAI has also been described differently by people involved. CEO Amjad Masad said at TechCrunch Disrupt on October 16 that the company was no longer working with Anthropic and that ChatGPT was better. He later told TechCrunch that Replit remained model-agnostic and that OpenAI was its preferred partner. A person familiar with the partnership told TechCrunch it remained in place. Those statements point to a changing supplier relationship, not a confirmed exclusive arrangement.

Deal Terms Remain Unpublished

The full terms of OpenAI’s reported customer offers are not public. The source does not specify how long the discounts last, whether they depend on usage or exclusivity, or how the bank’s additional seats are governed. The Information’s story is paywalled, and the source material relies in part on its public summary and other outlets’ accounts.

It is also unclear whether the reported pricing changes will produce lasting customer shifts. Rippling’s figures are a snapshot of one company’s coding work, and Replit’s public comments and outside accounts do not settle the precise status or terms of its vendor relationships. Nor do list prices establish what customers pay after credits, cached-token charges, negotiated rates and usage limits are applied.

Finally, the quoted savings estimates are not independently audited in the material provided. Model performance may differ by task, and the price comparisons involve successor models as well as rate changes. The available information does not establish that lower prices will translate into equivalent quality or lower total spending for every buyer.

Buyers Compare Offers and Limits

Companies considering a renewal or vendor change can compare written offers alongside measured performance on their own workloads. That means checking token rates, cache pricing, included usage, overage charges, credits and expiration dates, as well as any seat minimums or commitments. The reported deals make negotiation relevant, but there is no confirmed timetable for further OpenAI or Anthropic price changes.

Customers will also be watching whether reported discounts become more widely available and whether usage shifts persist at companies such as Rippling and Replit. For subscribers, the immediate practical step is to review current plan allowances: a monthly price can stay the same even when included usage changes. Further details will depend on new vendor announcements, published pricing pages and disclosures from customers with negotiated contracts.

Key Questions

Is OpenAI offering a 90% discount to all customers?

No. The 80% to 90% figure refers to DataCamp CEO Jonathan Cornelissen’s estimate of the total savings his company expects from OpenAI’s offer compared with its existing setup. It is not a published, general discount from OpenAI’s list prices.

How much did the listed model prices change?

According to the source, OpenAI’s GPT-6.1 Sol is listed at $2 per million input tokens and $10 per million output tokens, compared with an earlier Sol price of $5 and $30. Anthropic’s Opus 5.5 is listed at $4 and $20, down from $5 and $25 for Opus 5. These comparisons involve different model releases.

Are companies switching from Anthropic to OpenAI?

Some reported customer usage has shifted, including at Rippling, where CEO Parker Conrad said OpenAI’s share of coding work increased. The available accounts do not show that all customers are switching or that the trend applies across the market.

What should AI buyers compare besides token prices?

Buyers should check included usage, cache rates, credits and expiration dates, seat minimums, overage charges and contract commitments. They should also test model performance on their own work, because lower prices across successor models do not by themselves establish equivalent results.

Source: ThorstenMeyerAI.com