Google has introduced updates to its Gemini family of large language models, unveiling a new product line with competitive pricing aimed at low-cost AI models such as those from Chinese rival DeepSeek.
The company already offered a lightweight version known as “Flash,” but it has now launched “Flash-Lite,” an even cheaper alternative. As part of the updates, Google has made Gemini 2.0 Flash available to the general public after previewing it to developers in December, while also introducing Flash-Lite and releasing a test version of its flagship “Pro” model.
Koray Kavukcuoglu, chief technology officer of Google’s DeepMind AI lab, stated that Flash-Lite was developed following positive feedback on the 1.5 version of Flash.
Meanwhile, Gemini 2.0 Flash comes at a higher cost than its predecessor. The pricing and cost structures of AI models have drawn increased investor attention, particularly after DeepSeek disclosed it spent under $6 million on the final training run of a model — a significantly lower figure than what leading U.S. AI firms are estimated to spend.
DeepSeek’s emergence has also sparked discussions in earnings calls from Alphabet, Microsoft, and Meta, all of which remain committed to substantial capital expenditures in AI, Reuters news report said.
In terms of cost, certain inputs for Gemini Flash-Lite are priced at $0.019 per 1 million tokens, compared to $0.075 for the cost-efficient version of OpenAI’s flagship model and $0.014 for DeepSeek’s low-cost model. However, DeepSeek has stated that its pricing will increase fivefold on February 8, which could shift the competitive landscape.
InfotechLead.com News Desk
