San Francisco:
Alphabet Google on Wednesday announced updates of its GEMINI family of large language models, including a new line of products with competitive prices to low -cost artificial intelligence models such as that of the Chinese rival Deepseek.
The technological giant offers several Gemini versions that vary in price and performance. He already offered a light variant known as “Flash”, but his new “Flash-Lite” model is even cheaper.
On Wednesday, Google launched Gemini 2.0 Flash to the general public after obtaining a view prior to developers in December. He also launched Flash-Lite and launched a new version of his “Pro” Model in test phases.
Google created Flash-Lite after receiving positive comments on Flash version 1.5, Koray Kavukcuoglu, Director of Laboratory Technology of Google’s Deepmind, said in a press release. Gemini 2.0 Flash is more expensive than its predecessor.
The cost of developing AI models and, in turn, the cost of using them has been under the scrutiny of investors in recent weeks after Depseek revealed that he spent less than $ 6 million in the final execution of a model of a model of a model. The developers of the main companies of the United States said that the total cost was probably the largest magnitudes.
Even so, Depseek’s rise asked questions about Alphabet’s earnings calls and Microsoft and Meta rivals. Until now, everyone has indicated the intention of continuing mass capital expenses in the field.
Alphabet’s shares fell on Tuesday in part due to the pessimism of investors around a planned Capex increase that was 29% higher than Wall Street expected.
Certain entries in Gemini Flash-Lite cost $ 0.019 for 1 million tokens, a term for data units processed by an AI model. That is compared to $ 0.075 in the profitable version of the Openai flagship model and $ 0.014 in the Cheap model of Deepseek, although Deepseek states on its website that the price will increase five times on February 8.
(This story has not been edited by NDTV staff and was automatically generated from a union feed).
By
