"Let's Slow Down AI," They Said—Then OpenAI and Anthropic Hit Back With Cheaper New Models

[Financial News, New York—Reporter Lee Byung-chul]  Amid growing calls to slow the pace of artificial intelligence (AI) development, OpenAI and Anthropic have unveiled new, more affordable AI models. Although voices within the industry are calling for a slower pace of advanced-model development to improve AI safety, the market is seeing even fiercer competition over price and efficiency as companies seek to compete with China's low-cost open-weight models.
According to CNBC and other outlets on the 22nd, local time, OpenAI added GPT-6 Sol and GPT-6 Luna to its GPT-6 lineup. API prices were cut by 50% compared with the promotional pricing for GPT-5.6.
GPT-6 Sol is one tier below the top-of-the-line model, Astra, and focuses on coding and other complex tasks. GPT-6 Luna was designed to quickly handle large volumes of work, such as information extraction and document summarization. Rather than pursuing maximum performance across the board, OpenAI is segmenting performance and costs by task to reduce AI expenses for corporate customers.
Anthropic also unveiled Claude Opus 5.5, the latest model in its Claude 5.5 lineup. The company explained that improved token efficiency had lowered operating costs by about 40% compared with the existing Opus 5.
Dianne Penn, Anthropic's head of product management, research, and labs, said, "One of the areas in which we continue to innovate is how to make thinking and responses more efficient." She explained that users can reduce token usage depending on the level of reasoning they select.

Separate from the recent safety debate, competitive pressure in the market is intensifying. As corporate customers feel the burden of rising AI spending, they are demanding models that can perform the same tasks at a lower cost.
Chinese companies such as Alibaba Group, Moonshot AI, and DeepSeek have been rolling out relatively inexpensive open-weight models one after another. As a result, leading U.S. AI companies are under pressure to lower prices and improve token efficiency.
The competitive focus of the AI industry is expanding beyond simply asking, "Who can build the most powerful model?" to asking, "Who can deliver the highest performance at the lowest cost?" While calls to slow AI development are growing, the market continues to compete to produce cheaper and more efficient models.




[email protected] Reporter Lee Byung-chul Reporter