OpenAI has significantly reduced prices on two of its ChatGPT-5.6 models, a move that signals an intensifying price war in the AI model market and comes as enterprises increasingly scrutinize their AI expenditures.

A July 30 announcement on the OpenAI website confirmed that the cost of GPT-5.6 Luna — the company's fastest and most affordable model — will drop by 80%, while the mid-tier GPT-5.6 Terra, designed for everyday workloads, will see a 20% reduction. The price of the flagship Sol model remains unchanged, though OpenAI says it is now faster in the API. All three models were introduced three weeks ago.

"These updates help customers get more from every dollar they invest in AI and move faster when time matters," OpenAI stated.

The new API pricing is now in effect. Terra is priced at $2 per million input tokens and $12 per million output tokens, while Luna costs $0.20 per million input tokens and $1.20 per million output tokens. For comparison, Anthropic's mid-tier Claude Sonnet 4.6 is priced at $3 per million input tokens and $15 per million output tokens. Because API customers — primarily developers and enterprises building applications — pay based on the volume of text processed, per-token price changes translate directly into the unit economics of AI-powered products and services.

Additionally, a new Fast mode for GPT‑5.6 Sol replaces Priority Processing in the API. OpenAI claims Fast mode operates at speeds more than two-and-a-half times faster than standard processing.

Efficiency Gains Enable Reductions

OpenAI attributed the price cuts to efficiency improvements achieved during the development of its GPT-5.6 foundation model.

"Our strategy remains focused on advancing both capability and efficiency so each generation of intelligence can accomplish more work at a lower cost," the company said in a statement. Industry-wide, providers have been driving down inference costs through a combination of model architecture refinements, quantization techniques, and optimized serving infrastructure.

Enterprises Questioning AI ROI

The reductions arrive as organizations grow increasingly concerned about the cost of AI adoption. A recent vendor survey found that enterprise AI spending has grown rapidly with little accountability, and many companies remain uncertain whether their investments are delivering returns.

In June, Uber drew attention when it placed a per-employee cap on AI expenditure, highlighting the growing pressure to rein in costs.

A Crowded and Competitive Market

The pricing adjustments also reflect a fiercely competitive landscape. Lower-cost Chinese models from vendors such as Alibaba and Moonshot are competing directly with offerings from U.S. technology leaders including Google and Microsoft. As leading frontier models increasingly converge in benchmark performance, providers are differentiating more aggressively on price, speed, and total cost of ownership. The cuts further underscore the rivalry between OpenAI and Anthropic, both of which are proceeding toward public offerings — a timeline that adds pressure to demonstrate revenue growth and enterprise traction to prospective investors.

OpenAI's decision marks the latest salvo in what is shaping up to be a sustained price war, as model providers vie for enterprise customers who are simultaneously demanding stronger evidence of return on their AI investments.