OpenAI Reduces Prices for GPT-5.6 Terra and Luna Models

OpenAI’s GPT-5.6 series, launched earlier this year, includes the flagship Sol model, the mid-tier Terra, and the fast-response Luna. Enterprises have expressed reluctance to adopt the most capable models because of high usage fees measured in “tokens.” The company also faces mounting competition from Chinese startups such as Z.ai, as well as from Google, Microsoft and Anthropic, whose offerings are positioned at higher price points.

CNBC

On July 30, OpenAI announced that the price of the Terra model will drop 20 % to $2 per million input tokens and $12 per million output tokens, down from $2.50 and $15 respectively. The Luna model’s cost is being cut 80 % to 20 cents per million input tokens and $1.20 per million output tokens, down from $1 and $6. Sol’s pricing remains unchanged. By contrast, Anthropic’s Claude Sonnet 4.6 mid-tier model costs $3 per million input tokens and $15 per million output tokens, placing it above the new Terra rates. OpenAI attributes the ability to lower prices to efficiency gains built into the GPT-5.6 architecture, which it says improve code and performance during internal development.

Reuters

OpenAI framed the reductions as part of a broader effort to make each generation of its models “more capable and more efficient,” aiming to enable businesses to perform tasks that previously required top-tier systems at a substantially lower cost. The company emphasized that the price cuts target the low- and mid-tier models most commonly used by enterprises, with the expectation that reduced token fees will encourage broader adoption across cost-sensitive customers.

Reuters

• “Our strategy remains focused on advancing both capability and efficiency so each generation of intelligence can accomplish more work at a lower cost,” — Terra. OpenAI

CNBC