A Closer Look At The AI Deals OpenAI And Anthropic Are Offering

🔍 Read the full analysis: A Closer Look At The AI Deals OpenAI And Anthropic Are Offering on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get everyday essentials 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, while OpenAI is also offering some customers individually negotiated deals, according to vendor pricing pages and reporting by The Information. The moves matter to businesses buying AI services, but the widely cited claim of 80% to 90% savings describes one buyer’s expected total cost reduction, not a general discount off OpenAI’s list prices.

A comparison of vendor prices retrieved October 26 shows 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. Its smaller Luna model fell from $1 and $6 to $0.10 and $0.50. These are prices for different model generations, not simply reductions to an unchanged product.

Anthropic also lowered prices across model releases. Opus 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. Anthropic also reduced cache-read prices for some models. The source report says cached tokens can account for a large share of spending on coding agents, which repeatedly refer to conversation history; the practical bill can depend on this usage as well as headline input and output rates.

Separate from list prices, The Information reported customer-specific offers from OpenAI. One person cited in its reporting said Replit received a discount of more than 30% on a contract worth over $100 million annually. Another reported offer gave a U.S. bank, described as having more than $7 billion in assets, 500 ChatGPT seats at no additional charge. The source material does not provide the bank’s name or full contract terms.

DataCamp CEO Jonathan Cornelissen said OpenAI’s pricing could reduce his company’s costs by 80% to 90%, according to the cited reporting. That figure is an expected saving against the company’s existing setup, not a discount of that size from OpenAI’s published price list. The source material does not specify the full cost calculation or which services are included.

How AI Buyers Can Read the Offers

Falling prices and negotiated offers can reduce the cost of using AI for software development and other business tasks. They also give buyers more leverage when renewing contracts or comparing providers. But a headline price alone does not show what an organization will pay: model choice, token usage, cached input, seat counts, allowances and contract terms can all affect the total.

The reported customer shifts suggest price may influence which models companies use. Rippling CEO Parker Conrad said the share of his company’s coding work handled by OpenAI models rose from 16% in May to 52%, while Anthropic’s share fell from 57% to 18%. Those are figures about Rippling’s own work, not a market-wide measure, and Conrad attributed the change to cost.

For customers, the practical implication is to compare offers against actual workloads and to read the limits attached to a subscription. A lower rate may apply to a newer model, a specific term, or a particular usage pattern; it does not by itself establish that the model will perform equally well on a buyer’s tasks.

Price Cuts and Contract Fine Print

The price sequence compiled in the source report spans several model launches and adjustments from June through September 2026. It draws on pricing announcements and vendor pages, with a comparison to figures tracked by data scientist Max Woolf. In June, OpenAI previewed GPT-5.6 Sol and Luna at $5/$30 and $1/$6 per million input/output tokens. Anthropic launched Sonnet 5 at introductory pricing of $2/$10 in June and later made that price permanent, according to the source.

OpenAI cut Luna prices in July and Sol prices in August before introducing GPT-6 Sol and Luna in September. Anthropic released Opus 5 in July and Opus 5.5 in September. The comparison therefore includes successor models as well as price changes. OpenAI described GPT-6.1 Sol as offering “near-Astra intelligence for a fifth of the price,” but whether a newer model performs as well on a particular customer’s work requires testing on that work.

Subscription terms have shifted alongside API prices. The source report says Anthropic added capped, expiring API credits to its Max and Team plans, and that its Enterprise plan requires 20 seats. It also says OpenAI reduced the usage included with new $200 Pro subscriptions without changing the monthly price. These details come from the supplied report; readers should verify current terms directly with each provider before purchasing.

What the Deal Figures Leave Open

The full terms of the reported OpenAI offers are not public in the supplied material. It does not identify the U.S. bank, spell out the conditions attached to its 500 seats, or show how long the reported Replit discount lasts. The precise basis for DataCamp’s projected 80% to 90% cost reduction is also not detailed.

Some figures in the reporting are attributed to people familiar with deals, rather than published contracts. Replit’s relationship with OpenAI is described differently across accounts: Masad called OpenAI its preferred partner, while a person familiar with the partnership told TechCrunch it remained in place. Replit said it remained model agnostic. The material does not establish the precise scope or exclusivity of the arrangement.

It is also not clear from list prices alone whether newer models will deliver comparable results for each buyer, or how customer bills will change under revised credits, allowances and seat requirements. The listed token prices and subscription terms can change; current terms should be checked with the vendors.

What Buyers Should Check Next

Companies considering a renewal or a switch can request written quotes from multiple providers and compare them using representative workloads. That comparison should include input and output tokens, cached-token rates, seat minimums, usage caps, credit expiry and renewal terms, not only the advertised model price. Testing model quality on the same tasks can help separate a nominal price cut from a change that also affects performance.

Further developments will depend on whether the reported offers are extended to more customers and whether vendors make their terms public. Buyers should also monitor changes to subscription allowances and API credits, since the monthly fee may stay the same even when included usage changes. No additional deal details or next pricing announcement are confirmed in the supplied material.

Key Questions

Does the reported 80% to 90% saving mean OpenAI cut list prices by that amount?

No. The figure attributed to DataCamp CEO Jonathan Cornelissen describes expected total savings against DataCamp’s existing setup. It is not a general discount from OpenAI’s published prices, and the full calculation was not provided.

Which OpenAI model prices fell?

According to the source report’s comparison of vendor prices, Sol moved from $5/$30 to $2/$10 per million input/output tokens between June and September, while Luna moved from $1/$6 to $0.10/$0.50. The figures involve different model generations.

What customer-specific OpenAI deals were reported?

The source material describes a reported discount of more than 30% for Replit on a contract worth over $100 million a year, and a U.S. bank receiving 500 ChatGPT seats at no extra charge. Full terms were not disclosed.

Are Anthropic’s prices also falling?

Yes. The source report says Opus pricing moved from $5/$25 per million input/output tokens for Opus 5 to $4/$20 for Opus 5.5. It also describes lower cache-read prices for certain releases.

What should businesses compare before choosing a plan?

Businesses should compare total costs for their own workloads, including model performance, token use, cache rates, included usage, seat requirements, credits and contract conditions. Published list prices do not disclose every negotiated term.

Source: ThorstenMeyerAI.com