Anthropik, the emerging company backed by Google and Amazon, revealed on Monday the third series of artificial intelligence models known as “Cloud” in the latest rounds of semi-continuous competition in Silicon Valley to release more efficient technology, according to Reuters.
The startup company says that “Cloud 3 Opus” – the most advanced version in the series – outperforms in many benchmark tests the competing models ChatGPT 4.0 developed by OpenAI and Gemini 1.0 by Google.
The company’s CEO, Dario Amodei, said: “These are Rolls Royce artificial intelligence models (referring to the luxury car model and the best performance), at least for now.” The preview of the new or expected advanced models from Anthropik was unable to be verified, according to Reuters.
Following countless competitive announcements, as well as the release of the Cloud 2.0 series in July last year, these developments show how companies are competing to develop the performance of artificial intelligence tools while customers continue to explore possible ways to use this technology.
Daniella Amodei, President of Anthropik, said that customers will choose the new series despite the high price “if they need to perform tasks more cognitively complex”, such as dealing with precise financial analysis.
The company explained that the Opus version within the third series will cost $15 per million units of data, and the cost of the least capable models will be reduced by at least five times for dealing with the same amount of data.
On the other hand, OpenAI charges ten dollars per million data units dealt with by its “GPT 4 Turbo” model. Anthropik also stated that “Cloud 3” is the first “multi-modal” artificial intelligence series to be launched. This means – like some other competing models – that the model can respond to text and image queries, for example, analyzing an image or table.
This artificial intelligence model, which is “generative” and which was instrumental in the spread of “ChatGPT” application, can create new content on demand.
The company’s CEO, Dario Amodei, said that “Cloud 3” will examine images but will not produce them because we have “seen much less demand from institutions for them”.
Google temporarily suspended the image creation feature in its chatbot Gemini last month after it failed to meet diversity requirements, such as inaccurately depicting Nazi soldiers in the 1940s as having different racial backgrounds instead of being white-skinned.
Anthropik said that the “Cloud 3” series will be available through Amazon and Google cloud platforms, and added that it will also provide direct access in 159 countries.
Leave a Reply