AI price war: OpenAI closes the gap with Anthropic in the enterprise segment

Anthropic took the lead in enterprise AI this year, but that lead has little cushion: over the past few months OpenAI has noticeably narrowed the gap. The source material gives no market shares, prices, or contract volumes. What it does offer is a frame: a market that recently argued over benchmarks is increasingly arguing over price.

What put Anthropic ahead

There are two reasons. The first is its frontier models, which companies judged strong enough for real work. The second, and possibly the more important one, is its service for writing code. According to the source, American businesses rated it highly.

Coding has become an ideal entry point for enterprise sales. The result is visible right away: a development team either closes tasks faster or it doesn't. Token consumption is steady, and it grows with the autonomy of agents, which now work on a task for hours rather than seconds. A company that gets a foothold in engineering departments gains access to the rest: support, analytics, legal, internal bots.

What has changed at OpenAI

OpenAI has narrowed the gap in the enterprise segment. The source does not say how, and it would be dishonest to guess at specific prices or deals. But the logic of this kind of catch-up is familiar. OpenAI has huge ChatGPT name recognition among employees, ready-made enterprise products, and its own developer tools. To turn that into contracts, it has to remove the buyer's main objection: why should we pay more?

That explains the focus on price in the source's headline. When the flagship models of the two leaders are close in quality on typical tasks, total cost of ownership decides: price per token, volume discounts, limits, support terms.

What is known and what is not

• Leader of the year: Anthropic, thanks to its frontier models and its coding service.
• Challenger: OpenAI, which has narrowed the gap specifically in the enterprise segment.
• Main lever of competition: price, judging by the source's interpretation.
• Not disclosed: market shares, per-million-token rates, the size of the largest contracts, quarter-by-quarter trends.

The last point matters. Without numbers, it is impossible to say whether this is a small shift in OpenAI's favor or a change of leader. For now it is a trend, not a confirmed turning point.

How to choose a model in this situation

Price competition benefits the buyer, but only if the buyer is ready to take advantage of it. A few practical steps.

• Don't lock in with a single vendor for the long term. In six months the terms may be noticeably better.
• Build an abstraction layer over model APIs so that switching takes days, not months.
• Compare the cost of a solved task, not price lists. An expensive model that gets it right on the first attempt often turns out cheaper.
• For coding, measure on your own repositories: share of accepted changes, review time, tokens spent per task.
• Before renewing a contract, ask for the terms to be recalculated. While the leaders are fighting for enterprise customers, the buyer's negotiating position is stronger than usual.

The main question for the coming quarters is whether Anthropic can hold its advantage in coding if OpenAI keeps pressing on price. Revenue reports and the terms of new enterprise deals will answer it, and those are what to watch.