OpenAI Announces 80% ChatGPT Price Cut Amid Kimi K3’s Breakthrough Success

OpenAI has lowered ChatGPT API pricing by up to 80%, with GPT-5.6 Luna input costs dropping to $0.20 per million tokens. The reductions come as competition from China's Kimi K3 intensifies.

OpenAI announces 80% ChatGPT price cut amid Kimi K3’s breakthrough success
OpenAI announces 80% ChatGPT price cut amid Kimi K3’s breakthrough success

OpenAI has reduced the price of its most popular AI models. Two models from the GPT-5.6 family, the GPT-5.6 Luna and the GPT-5.6 Terra, are being discounted by OpenAI, according to CEO Sam Altman. This decision comes as OpenAI is confronted with fierce competition from more affordable open-weight models developed in China, such as Kimi K3.

The news was broken by Sam Altman on X. He announced the price drops in a post. "GPT-5.6 Luna has dropped 80 percent, to $0.20 per million input tokens and $1.20 per million output," according to his post. GPT-5.6 Terra prices fall 20% to $2/$12.

OpenAI Waging Price Battle in AI Sector

Luna and Terra are now less expensive than the biggest open-weight AI model available, Kimi K3, thanks to pricing reduction. Kimi K3 has an input token price of $3 (about INR 287) and an output token price of $15 (about INR 1,436) per million. After its release, Kimi K3 stunned the artificial intelligence (AI) world with its ability to compete with GPT-5.6 Sol and Anthropic's Claude Fable 5 on specific tasks—all while costing a fraction of the price. However, OpenAI is not focusing solely on expenses.

The AI business has also released a rapid mode for its most sophisticated AI model, GPT-5.6 Sol, for individuals who desire top-notch performance quickly. According to Sam Altman, GPT-5.6 Sol has a Fast option in the API, which allows for 2.5 times the speed at twice the price, while maintaining the same smarts. To rephrase, Sol offers far quicker responses to users, but at twice the price. "Our goal is to provide the optimal balance between price and intelligence across all levels," Altman emphasised.

AI Companies Now Focusing on Cost-Cutting

OpenAI is also working on improving the efficiency of its models. The AI startup claims to have improved its operational efficiency following deployment by utilising GPT-5.6 Sol. As a result, service expenses were cut by 20%, and token-generation efficiency was improved by over 15%. Companies are becoming more price conscious and are hesitant to engage in costly AI systems without a solid plan on how to recoup those costs.

This has led to a general shift in pricing. Companies like Uber and Walmart are attempting to rein on spending on artificial intelligence as the costs have skyrocketed, sometimes reaching into the billions. The Chinese open-weight competitors are putting pressure on OpenAI and other US AI institutes as well. These rivals may be downloaded, tweaked, and deployed on a company's own infrastructure.

For example, Kimi K3's weights are now freely downloadable. Consequently, programmers will be able to optimise the model by running it locally. OpenAI's new pricing structure puts greater pressure on competitors like Google and Anthropic, who have all been promoting cheaper or more cost-effective AI models as of late. New, less expensive models have also been released by Meta and xAI.