Alphabet’s Google has unveiled a new class of budget-friendly AI models in its Gemini family, responding to growing concerns over the cost of AI development and usage. The latest addition, Gemini Flash-Lite, aims to compete with low-cost AI models, including those from Chinese rival DeepSeek.
Google Expands Gemini AI Lineup
Google offers multiple versions of its Gemini AI models, varying in price and performance. The company already had a lightweight version called Gemini Flash, but the newly introduced Flash-Lite is an even cheaper alternative.
Alongside Flash-Lite, Google also:
- Released Gemini 2.0 Flash for public use after a developer preview in December.
- Launched a new version of its flagship “Pro” model for testing.
Rising Cost Concerns in AI Development
The cost of developing AI models has come under intense investor scrutiny, especially after DeepSeek revealed it spent less than $6 million on training a model—significantly lower than industry estimates for major US firms.
Investors have raised concerns about capital expenditures in AI, leading to Alphabet shares dropping on Tuesday following a planned 29% higher spending hike than expected.
Competitive AI Pricing: Google vs. DeepSeek vs. OpenAI
The cost of using AI models varies significantly:
- Gemini Flash-Lite charges $0.019 per 1 million tokens.
- OpenAI’s cost-efficient model is priced at $0.075 per 1 million tokens.
- DeepSeek’s budget model costs $0.014 per 1 million tokens, but pricing is set to increase fivefold on February 8.
Conclusion
With the AI industry rapidly evolving, Google’s new Gemini models aim to make AI more accessible and cost-effective, while competitors like DeepSeek and OpenAI continue to push their pricing strategies. However, rising investor concerns over AI development costs could shape the industry’s future financial decisions.
