alphabet will release three new Gemini models on Tuesday. This contains the clearest answer so far. humanThe company has become a leader in the cybersecurity space as it seeks to show progress across its product pipeline in the face of delays and increased competition.
Gemini 3.5 Flash Cyber is designed to detect and patch software vulnerabilities and will initially be available only to governments and trusted partners through a limited access pilot. Google said specialized models will have a lower price per token than large-scale models.
This could help Google close the cybersecurity gap with Anthropic, which established an early lead in automated code protection.
Google will also launch Gemini 3.6 Flash. This improves performance for coding, multimodal, and knowledge work, reduces token usage by up to 17% over previous models, and reduces cost per token. This significantly reduces the cost of running high-volume workloads.
Gemini 3.5 Flash-Lite, on the other hand, is the fastest and cheapest model in Google’s 3.5 family, built for high-volume workloads and small tasks within large-scale AI agent systems.
The broad lineup reflects Google’s bet that price and efficiency can offset timing delays in some key product categories.
Artificial analysis data shows that the Gemini Flash is already underperforming its Anthropic counterpart. OpenAI And it is a rival to China in terms of cost. The company says Gemini 3.6 Flash, the more powerful of the two new models, costs less per task than GPT-5.6 Terra Max, Kimi K3, and Qwen 3.7 Max, and 3.5 Flash-Lite costs a fraction of that.
The development comes on the eve of Alphabet’s earnings results and comes as Chinese rivals gain momentum. Moonshot AI’s Kim K3 attracted enough demand that the company limited new subscriptions and API access due to capacity constraints, while Alibaba teases Qwen 3.8 Max, saying it will only rival Anthropic’s Fable 5 in overall performance.
This demand highlights the other side of the AI race. So building a competitive model is only part of the challenge. Businesses also require sufficient computing power to provide services at scale.
Although Google faces its own capacity constraints, it has potential advantages through custom chips, cloud infrastructure, and the ability to co-design models and hardware.
Tuesday’s model announcement comes as Google is reportedly developing a specialized chip designed to run Gemini up to 10 times more efficiently, as part of a broader effort to reduce the cost of its AI services.
A Google Cloud spokesperson told CNBC that the company’s teams are “constantly researching and experimenting with new innovations to deliver maximum performance and efficiency for our users and customers.” “While not every project will make it to production, this rigorous exploration is central to our full-stack approach,” the company said in a statement.
“Co-designing the hardware and software from the ground up ensures that our systems are integrated and highly optimized for real-world workloads,” the statement continued.
Google is also providing more visibility into its roadmap following questions about delays. Gemini 3.5 Pro is currently being tested with partners ahead of wide availability, while the company has begun its largest-ever pre-training for Gemini 4.
Attention: Downo record approaches with Alphabet as new driver
