ChatGPT developer OpenAI has cut the cost of its GPT-5.6 artificial intelligence (AI) models by as much as 80% and introduced faster processing for its premium model.
This, as competition intensifies among AI providers to make enterprise AI more affordable and scalable.
The company announced yesterday that GPT-5.6 Luna – its fastest and lowest-cost model – is now 80% cheaper, while the price of GPT-5.6 Terra, its general-purpose enterprise model, has been reduced by 20%.
It notes that the changes take immediate effect across the OpenAI API, ChatGPT Work and Codex platforms.
The move follows OpenAI’s disclosure a day earlier that GPT-5.6 had helped optimise parts of its own production infrastructure, enabling the company to reduce the cost of running its models. OpenAI said it is now passing those efficiency gains directly to customers.
“Making advanced intelligence more abundant and affordable is central to OpenAI’s mission to ensure AGI [artificial generative intelligence] benefits all of humanity,” the company says.
The pricing changes come as AI vendors increasingly compete on cost, speed and performance.
OpenAI’s latest announcement also introduces “fast mode” for GPT-5.6 Sol, its highest-performing model available through the API.
According to the company, the new mode delivers responses up to 2.5 times faster than the standard service for twice the processing cost.
It notes that under the new pricing structure, GPT-5.6 Terra will cost $2 per million input tokens and $12 per million output tokens, while GPT-5.6 Luna drops to $0.20 per million input tokens and $1.20 per million output tokens. Pricing for GPT-5.6 Sol remains unchanged.
According to OpenAI, the lower costs are made possible through improvements across its AI stack, including more efficient model routing, better inference software and smarter context management that reduces unnecessary computation.
The company also revealed that GPT-5.6 Sol has been assisting engineers in improving OpenAI’s own infrastructure.
Working within a human-supervised process, it says, the model rewrote production kernels, designed and executed hundreds of optimisation experiments, and monitored training systems for issues.
OpenAI says these improvements reduced the end-to-end cost of serving GPT-5.6 by 20%, while increasing token-generation efficiency by more than 15%.
The announcement comes amid an intensifying price war across the AI industry. Over the past year, major AI developers − including Anthropic, Google, xAI and Chinese competitors such as DeepSeek − have steadily reduced inference costs while improving model performance.
Falling prices have become a key battleground as vendors seek to drive enterprise adoption and encourage customers to shift more business processes onto AI platforms.
The updated pricing took effect on 30 July through the OpenAI API, with AWS rollout beginning the same day.
ChatGPT subscription prices remain unchanged, although users of ChatGPT Work and Codex will consume fewer usage credits when running GPT-5.6 Terra and Luna.
Share
