Google expands Gemini lineup with cheaper models and new Mythos rival - CNBC

What the report says

CNBC reported that Alphabet’s Google is adding three models to its Gemini artificial intelligence lineup, a move aimed at improving cost, speed and specialized capabilities as competition intensifies. The rollout, announced Tuesday, includes Gemini 3.5 Flash Cyber, a cybersecurity-focused model intended to identify and repair software vulnerabilities. CNBC said access to that model will begin through a limited pilot for governments and trusted partners.

The launch also includes Gemini 3.6 Flash, which Google says improves performance on coding, multimodal tasks and knowledge work while using up to 17% fewer tokens and costing less per token than the prior version. A third model, Gemini 3.5 Flash-Lite, is described as the fastest and lowest-cost option in the Gemini 3.5 family, aimed at high-volume uses and smaller tasks inside AI-agent workflows.

According to CNBC, the cybersecurity product represents Google’s clearest response so far to Anthropic’s early position in automated code defense with its Mythos model. The broader update underscores Google’s effort to compete not only on model quality, but also on efficiency and the cost of serving AI at scale. CNBC cited Artificial Analysis data indicating Gemini Flash is already priced below comparable offerings from Anthropic, OpenAI and Chinese competitors.

The timing is notable because the rollout comes one day before Alphabet reports earnings and as Chinese AI companies gain attention. CNBC reported that Moonshot AI’s Kimi K3 faced demand strong enough to constrain new subscriptions and API access, while Alibaba has been promoting its next Qwen model. Google has also pointed to its cloud infrastructure, custom chips and hardware-software design as potential advantages, while acknowledging ongoing experimentation to improve performance and efficiency.