OpenAI has announced a significant price reduction for two models in the GPT-5.6 lineup, effective July 30. The rate for the Luna model was cut by 80%, and for Terra by 20%. This decision comes amid growing caution among corporate clients regarding spending on artificial intelligence.

Just three weeks after the launch of GPT-5.6, OpenAI is revising its pricing policy. The input cost for Luna has been reduced from $1 to $0.20 per 1 million tokens, and output from $6 to $1.20. For Terra, prices are now $2 and $12 per 1 million tokens, respectively, compared to $2.50 and $15 previously. The flagship Sol model has retained its original rates.

Market pressure and rising competition

This move is a direct response to changing demand patterns. The corporate sector, which previously actively experimented with AI technologies, now demands measurable returns on investment. Company CFOs are blocking budgets without clear justification of efficiency, creating powerful pricing pressure on all market players.

Competition from Chinese developers is especially noticeable. Models like Kimi K3 from Moonshot AI and GLM-5.2 from Z.ai offer comparable quality at significantly lower prices. Even Anthropic's mid-tier model, Claude Sonnet 4.6, now costs more than OpenAI's updated Terra. This is a signal that the price war in the AI segment is only gaining momentum.

Strategy ahead of public offering

OpenAI explains the price cuts not as a forced measure, but as a result of increased efficiency. Each new generation of models should accomplish more tasks at lower costs. However, the timing is no coincidence: both companies—OpenAI and Anthropic—are preparing for initial public offerings.

Dumping has a dual impact on IPO preparation. On one hand, broader audience reach makes the business more attractive to investors, who at the listing stage look at revenue dynamics and client base growth rather than net profit. On the other hand, reduced margins on core services could negatively affect financial statements if request volumes do not grow proportionally.

The key question now is whether demand growth can offset the decline in revenue per token unit. If not, financial metrics will suffer, significantly complicating the creation of an attractive prospectus for future shareholders. OpenAI's upcoming financial report should reveal the first results of this strategy.

My analysis: The price reduction is a classic scaling strategy before an IPO, but it carries significant risks. OpenAI is betting that demand elasticity will offset the loss of margins. However, amid fierce competition from cheaper Chinese counterparts, this move may prove to be only a temporary measure, postponing the inevitable restructuring of the business model.