Anthropic and OpenAI Launch Low-cost AI Models for First Time Amid Competition from Open-weight Models

Anthropic and OpenAI each announced new low-cost AI models on Tuesday (22nd), as the two frontier AI labs continue to face sour competition from cheaper open-weight rivals.

OpenAI added GPT-6 Sol and GPT-6 Luna tiers to the GPT-6 series, with API pricing cut by 50% compared with the promotional pricing of GPT-5.6.

Anthropic on Tuesday launched Claude Opus 5.5, the latest version in the Claude 5.5 model series. The company said the new version improved token usage efficiency, with operating costs about 40% lower than Opus 5.

Dianne Penn, Anthropic's head of product management, research and labs, told CNBC that one of the company's ongoing innovation directions is improving reasoning and response efficiency, allowing models to use fewer tokens under different computing intensity settings.

The separate launches mark the first releases since Anthropic CEO Dario Amodei recently called on the industry to slow the development of advanced AI models. The issue has drawn significant attention as debates surrounding model safety have boiled up in recent weeks.

Former Anthropic researcher Jacob Coxon posted on X on September 8 that he had resigned, warning that the two labs were gambling with people's lives, sparking further debate.

OpenAI CEO Sam Altman, as well as Tesla and Space Exploration Technologies Corp. (SPCX.US) CEO Elon Musk, also supported Amodei's call.

As OpenAI and Anthropic roll out new models, they are seeking to meet customer demand for more cost-effective models and better control of AI spending.

The two labs are also facing pressure from companies offering cheaper open-weight models, including Chinese firms BABA-W, Moonshot AI and DeepSeek.