Alphabet released three new Gemini models Tuesday, including a specialized cybersecurity variant that takes direct aim at Anthropic’s early edge in automated code defense. The move comes as Google works to demonstrate momentum across a product pipeline that has faced delays and intensifying competition from Chinese rivals.
The standout release is Gemini 3.5 Flash Cyber, a model built to detect and patch software vulnerabilities. It will initially be available only to governments and trusted partners in a limited pilot. Google says the model costs less per token than larger alternatives, potentially narrowing the gap with Anthropic’s lead in cybersecurity. Alongside it, Gemini 3.6 Flash improves coding, multimodal, and knowledge-work performance while using up to 17% fewer tokens and lowering per-token costs—a meaningful reduction for high-volume workloads. The third model, Gemini 3.5 Flash-Lite, is the fastest and cheapest in the 3.5 family, designed for small tasks within larger AI-agent systems.
Google is betting that price and efficiency can offset its slower timing in key categories. According to Artificial Analysis data, Gemini Flash already undercuts comparable models from Anthropic, OpenAI, and Chinese rivals. Gemini 3.6 Flash is cheaper per task than GPT-5.6 Terra Max, Kimi K3, and Qwen 3.7 Max, while 3.5 Flash-Lite costs a fraction of that. The rollout comes as Chinese competitors gain traction: Moonshot AI’s Kimi K3 saw such high demand that it limited new subscriptions and API access, while Alibaba is teasing Qwen 3.8 Max, claiming it trails only Anthropic’s Fable 5. Google’s advantage lies in its custom chips and cloud infrastructure, though it has faced its own capacity constraints. The company is also reportedly developing a specialized chip to run Gemini up to 10 times more efficiently, part of a broader push to lower serving costs. A Google Cloud spokesperson said the company is “constantly researching and experimenting” with hardware-software co-design to optimize real-world workloads. Google is also offering more roadmap visibility: Gemini 3.5 Pro is now being tested with partners, and the company has started its largest-ever pre-training run for Gemini 4.
Source: CNBC