Google DeepMind releases Gemini 3.6 Flash and two new models to improve AI efficiency and cybersecurity.

Google DeepMind released three new AI models on Tuesday, focusing on cost-efficiency and specialized performance to compete with rivals like OpenAI and Anthropic. The primary release, Gemini 3.6 Flash, serves as a "workhorse" model that improves coding and multimodal capabilities while reducing token usage by 17% compared to its predecessor. This update makes the model cheaper for high-volume workloads, with costs set at $1.50 per million input tokens and $7.50 per million output tokens. The company also introduced Gemini 3.5 Flash-Lite, designed for high-throughput tasks and low-latency requirements, and Gemini 3.5 Flash Cyber. The latter is a specialized model fine-tuned to identify and repair cybersecurity vulnerabilities. Due to its "dual-use" nature, the administration announced that Gemini 3.5 Flash Cyber will be available exclusively to governments and trusted partners through a limited pilot program. While the flagship Gemini 3.5 Pro remains in testing following internal delays, the company has already begun its most ambitious pre-training run yet for the upcoming Gemini 4.

Sources