Google has launched Gemini 3.8 Flash, its third Flash model in just six weeks, with a focus on smarter agentic workflows. Gemini 3.8 Flash “works harder” on complex tasks, outperforming larger models on several benchmarks while remaining relatively inexpensive.

Gemini 3.8 Flash delivers strong Deep Sweep 1.1 benchmark results while keeping prices low. Google's new model generates 300 tokens per second.

Alphabet’s cadence problem finally has a price tag. CEO Sundar Pichai said Gemini 3.5 Pro would arrive in June, but it still hadn’t shipped as of September 2, 2026. What Google shipped instead was Gemini 3.

Gemini 3.8 Flash Cyber launches with frontier-level AI vulnerability patching and a 2.6x lead over commercial rivals in Google's Chrome Security evaluations -- but the model is restricted to governments,

Google Gemini 3.8 Flash launches as a coding-focused AI model that reportedly rivals Claude Opus 5 and GPT-5.6 Sol while being cheaper to run.

Google might not be at the AI frontier, but it’s made its way back to the Pareto frontier. Google has released Gemini