Google unveils its newest Gemini models, offering advanced reasoning, coding, and cybersecurity capabilities at competitive pricing.
Google has announced the launch of Gemini 3.8 Flash and Gemini 3.8 Flash Cyber, two new AI models designed to deliver faster reasoning, enhanced coding assistance, and specialized cybersecurity analysis—all at a price point aimed at broad adoption.
Key capabilities of Gemini 3.8 Flash
Gemini 3.8 Flash builds on the Gemini 1.5 architecture, offering improved token efficiency and lower latency for real‑time applications. The model excels in complex problem‑solving, multi‑step reasoning, and supports a wider range of programming languages, making it a versatile tool for developers and data scientists.
The model’s training incorporates a larger corpus of code repositories and technical documentation, which translates into more accurate code suggestions, bug detection, and automated refactoring suggestions.
Gemini 3.8 Flash Cyber: AI for security teams
Gemini 3.8 Flash Cyber is tailored for cybersecurity workloads. It can analyze logs, detect anomalous patterns, and generate remediation steps in natural language, helping analysts respond faster to threats.
The model also supports threat‑intel summarization, vulnerability prioritization, and can draft incident reports that align with industry standards such as MITRE ATT&CK.
Pricing and availability
Google positions both models as cost‑effective alternatives to competing offerings. Pricing is structured per token usage, with discounts for high‑volume customers and an introductory free tier for developers experimenting with the APIs.
- Pay‑as‑you‑go pricing for flexible scaling
- Free tier includes up to 1 million tokens per month
- Volume discounts for enterprise contracts
The models are available today through Google Cloud’s Vertex AI platform, and developers can access them via standard REST and gRPC endpoints.