Google debuts new class of cheap AI models after Deepseek raises cost questions

San Francisco – Alphabet’s (Googl, Googl) Google announced on Wednesday updates to its Gemini family of large language models, including a new product line with competitive prices at low cost artificial intelligence models such as Chinese rival Deepseek.

The technology giant offers several versions of Gemini that vary in price and performance. It already offered a lightweight variant known as “flash”, but its new “flash -lite” model is even cheaper.

On Wednesday, Google Gemini Released 2.0 Flash to the public after previewing it to developers in December. It also launched Flash -Lite and released a new version of its flagship “Pro” model for test phases.

Google created flash-lite after receiving positive feedback on the 1.5 version of Flash, Koray Kavukcuoglu, head technology officer for Google Deepmind Ai Lab, 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 under investor survey in recent weeks after Deepseek revealed it spent less than $ 6 million on the final training of a model. Developers at leading us AI companies said the total costs were probably sizes larger.

Still, Deepseek’s rise raised questions about earnings calls from the alphabet and rivals Microsoft (MSFT) and META (META). Everyone has so far signaled the intention of continuing massive capital costs in the field.

Alphabet shares fell partially on Tuesday due to investorpessimism about a planned CAPEX hiking trip 29% higher than Wall Street expected.

Certain inputs on Gemini Flash-Lite cost $ 0.019 per 1 million tokens, a term for the units of data processed by an AI model. It can be compared to $ 0.075 on the cost -effective version of Openai’s flagship model and $ 0.014 on Deepseeks cheap model, though Deepseek states on its website that pricing will rise five times on February 8.

(Reporting Kenrick Cai in San Francisco; Editing Christopher Cushing)