OpenAI cut prices on two of its GPT-5.6 models on July 30, 2026, slashing the cost of Luna by 80% and Terra by 20%. The move comes as businesses grow wary of ballooning AI bills and cheaper rivals chip away at the company's market share.
What the price cuts look like
Luna's input price dropped from $1 to $0.20 per million tokens. Output fell from $6 to $1.20. Terra's input went from $2.50 to $2, and output from $15 to $12. The company's flagship model, Sol, kept its price unchanged.
The reductions are steep, especially for Luna, which now costs a fifth of what it did. Terra's cut is more modest — about 20% on both ends.
Why now
AI costs have been climbing, and finance teams are starting to push back. A term has even emerged for the problem: 'tokenmaxxing,' where employees use AI freely without tracking what it costs. Companies that once rushed to deploy large language models are now asking harder questions about the return on investment.
OpenAI framed the cuts as part of a strategy to advance both capability and efficiency, not a defensive move. But the timing suggests pressure from all sides.
The competition
Cheaper AI rivals are closing in. Anthropic's Claude Sonnet 4.6, Moonshot AI's Kimi K3, and Zhipu AI's GLM-5.2 all offer competitive pricing. Even after the cuts, Claude Sonnet 4.6 still costs more per token than the discounted Terra. But the gap is narrowing, and the list of alternatives is growing.
OpenAI isn't the only one feeling the heat. Both OpenAI and Anthropic are eyeing potential IPOs, and thinner margins could complicate those plans.
Lower prices should make it easier for businesses to scale up their AI usage without shocking their budgets. But the cuts also squeeze OpenAI's own margins at a time when investors are watching closely. The company says the goal is to drive adoption through better cost efficiency, not to fight a price war.
Whether that strategy holds depends on how quickly the market responds — and whether the cheaper models can deliver the same results.
