Google expands Gemini lineup with cheaper models and new Mythos rival
Alphabet It launches three new Gemini models on Tuesday, including the clearest answer to the question so far. anthropicThe company is a leader in cybersecurity as it aims to advance its product line, which is facing delays and increased competition.
Gemini 3.5 Flash Cyber is designed to detect and patch vulnerabilities in software and will initially be available only to governments and trusted partners through a limited-access pilot. Google said the private model operates at a lower price per token than larger models.
The new model could help Google narrow the cybersecurity gap with Anthropic, which ranks first in automated code defense with its Mythos model.
Google is also releasing Gemini 3.6 Flash, which improves encoding, multimodal, and computing performance using up to 17% fewer tokens and at a lower cost per token than its predecessor; This means a significant reduction in the cost of running high-volume workloads.
Gemini 3.5 Flash-Lite, meanwhile, is Google’s fastest and cheapest model in the 3.5 family, designed for high-volume workloads and smaller tasks in larger AI agent systems.
The broader product mix reflects Google’s bet that price and efficiency can help offset slowing timing in several key product categories.
Synthetic Analysis data shows that the Gemini Flash currently undercuts Anthropic’s comparable models. OpenAI and Chinese competitors on cost. The most powerful of Alphabet’s new models, the Gemini 3.6 Flash, is cheaper per mission than the GPT-5.6 Terra Max, Kimi K3 and Qwen 3.7 Max, according to the company.
The app comes on the eve of Alphabet earnings and as Chinese rivals gain momentum. Moonshot AI’s Kimi K3 has attracted enough demand for the company limited new subscriptions and API access due to capacity constraints, Alibaba Making fun of Qwen 3.8 MaxIt is said to fall behind only Anthropic’s Fable 5 in overall performance.
This demand highlights another aspect of the AI race: Building a competitive model is only part of the challenge. Companies also need sufficient computing capacity to provide services at scale.
Although the company faces its own capacity constraints, Google has a potential advantage thanks to its custom chips, cloud infrastructure, and ability to co-design models and hardware.
Tuesday’s model launches come as Google reportedly developed a custom chip designed to run the Gemini up to 10 times more efficiently; this was part of a broader effort to lower the cost of serving AI.
A Google Cloud spokesperson told CNBC that its teams are “constantly researching and testing new innovations to deliver maximum performance and efficiency to our users and customers” and that “while not every project makes it to production, this rigorous research remains at the core of our full-stack approach.”
“By co-designing our hardware and software from the ground up, we ensure our systems are integrated and highly optimized for real-world workloads,” the statement continued.
Google is also offering more visibility to the roadmap following questions about delays. The company has begun its largest pre-training effort ever for Gemini 4, while Gemini 3.5 Pro is being tested with partners ahead of wider availability.
WRISTWATCH: Google will work on new AI chip to run models more efficiently: Report
