Anthropic, the company behind the Claude family of AI models, has developed a new hybrid model on the information reported Thursday.
Hybrid AI models use more computational resources for complex queries that require inference and quickly answer simpler tasks when needed. Humanity is expected to release it in the coming weeks, the report added.
Additionally, the model is said to allow customers to control the time it takes to resolve queries. As revealed in the report, developers can adjust the token using “slide scale.”
Another report said that humanity expects revenue to exceed $34 billion in 2027. In addition to the new technology, the company burned $5.6 billion in cash last year and is expected to halt cash burning by 2027.
Meanwhile, as CEO Sam Altman revealed, even Openai is taking a similar approach.
GPT-4.5 is internally called Orion, and will be the next release, not the last chain model.
Additionally, Openai plans to unify the O-Series and GPT-Series models, integrate all the tools, and enable a system that can determine the optimal thinking time for tasks. GPT-5 will be introduced with ChATGPT, and the API will be introduced as a system that incorporates a variety of OpenAI technologies, including O3. The company will not release the O3 as a standalone model.
Additionally, Anthropic CEO Dario Amodei revealed that the flagship Claude 3.5 Opus will be released in November last year.
Reports say Anthropic has finished training for the Claude 3.5 Opus.
Recently, humanity has published a report to showcase the impact of AI on how people work.
It was revealed that approximately 36% of all occupations use AI for at least a quarter of related tasks. Furthermore, 57% of use cases involved leveraging AI to enhance human capabilities, while 43% proposed automation. However, only 4% of occupations introduced AI usage in at least 75% of tasks.
It was also reported that Google has invested more than $1 billion in humanity in addition to a $2 billion round led by Lightspeed Venture Partners last month.
