The generative artificial intelligence landscape reached a pivotal milestone on September 22, with a synchronized pricing offensive from industry leaders Anthropic and OpenAI. By launching the Claude Opus 5.5 model and the GPT-6 Sol and Luna versions almost simultaneously, these tech giants have officially ushered in a new era: the democratization of costs. This strategy marks a paradigm shift, where the relentless pursuit of raw power is gradually giving way to an aggressive trade war aimed at securing the loyalty of enterprises, which are increasingly sensitive to their API consumption bills.
The figures illustrate the scale of these cuts. Anthropic has lowered the entry cost of its flagship model by 20% compared to the previous generation, while drastically optimizing cache management. As a result, users can expect a 40% reduction in their standard usage costs. For its part, OpenAI has struck back with prices slashed by half compared to the GPT-5.6 range. The GPT-6 Luna model has now established itself as an extremely competitive solution, featuring entry rates of $0.10 per million tokens, highlighting the labs' ambition to capture the massive data volumes processed by professional infrastructures.
This margin pressure is not solely the result of a rivalry between the two American companies; it is a response to a growing threat from China. The rise of open-weight models, which are often more accessible and customizable, is eating into significant market share. Faced with this agile international competition, OpenAI and Anthropic are seeking to safeguard their dominant positions by making their services indispensable and economically viable at scale. This is a critical strategic pivot intended to lock in corporate ecosystems before foreign alternatives become permanently entrenched in cloud infrastructures.
For Anthropic, this offensive comes during a crucial financial window, with an Initial Public Offering anticipated for this autumn. This market debut could prove to be one of the largest in tech history, making it essential to demonstrate robust growth and the ability to maintain margins while staying competitive. At the same time, these new models incorporate enhanced security protocols, addressing recent governance requirements and calls for more responsible AI, confirming that reliability remains a selling point just as decisive as price when it comes to attracting large organizations.