Key facts
- OpenAI reduced the price of its GPT-5.6 Luna AI model by 80%.
- OpenAI reduced the price of its GPT-5.6 Terra AI model by 20%.
- OpenAI kept the price of its Sol AI model at its current level.
- OpenAI updated ChatGPT to reject prompts mimicking authors' styles.
- The ChatGPT style mimicry restriction follows copyright lawsuits against AI companies.
- The ChatGPT style mimicry restriction follows a ruling against Anthropic.
- Nscale is acquiring AI cloud startup Anyscale.
- The acquisition price for Anyscale is approximately $1.65 billion.
- Anyscale's platform is built around open-source Project Ray.
- Project Ray helps companies scale AI workloads.
OpenAI has implemented price reductions for two of its AI models, GPT-5.6 Luna and Terra, cutting their costs by 80% and 20% respectively. The price for its flagship Sol model remains unchanged. This strategic pricing adjustment is designed to appeal to businesses that are increasingly scrutinizing their artificial intelligence expenditures. The move also serves to intensify competition within the AI market, particularly against rivals such as Anthropic and various Chinese open-source models.
In parallel, OpenAI has introduced a new restriction within ChatGPT, prohibiting the AI from generating content that mimics the writing styles of specific authors, whether they are living or deceased. This policy update comes in the wake of numerous copyright lawsuits filed against AI companies and a notable legal ruling that went against competitor Anthropic. The restriction aims to address concerns surrounding intellectual property and the unauthorized use of creative works.
Meanwhile, in a significant acquisition within the AI sector, AI cloud provider Nscale has announced its intention to acquire AI software startup Anyscale for approximately $1.65 billion. This acquisition is intended to enhance Nscale's capabilities in managing AI systems. Anyscale's platform is built upon the open-source Project Ray, a technology that assists companies in scaling their artificial intelligence workloads efficiently.