Google says coding agents often produce stale Gemini API code because model training data has a cutoff date, and is shipping Docs MCP plus Developer Skills as the fix. Used together, Google reports a 96.3% pass rate with 63% fewer tokens per correct answer than vanilla prompting on its eval set.
Google is adding Flex and Priority service tiers to the Gemini API so developers can choose lower-cost synchronous inference for background work or higher-assurance routing for critical traffic. The change gives agent builders a cleaner way to separate cost and reliability without splitting architectures across multiple APIs.
Google introduced Veo 3.1 Lite as its most cost-effective video generation model, priced at less than 50% of Veo 3.1 Fast while keeping the same speed. The model is rolling out through the paid tier of the Gemini API and Google AI Studio, broadening access to higher-volume video app use cases.
Google AI said on March 25, 2026 that Lyria 3 Pro is now available across a broad mix of consumer, developer, and enterprise surfaces. The rollout suggests Google wants music generation to become part of its mainstream AI stack rather than a standalone experiment.
A high-ranking Hacker News thread amplified a Truffle Security report arguing that legacy Google API keys can become high-impact credentials when Gemini APIs are enabled. The post highlights exposure scale claims and concrete key-hardening steps.
On 2026-02-19, Google announced Gemini 3.1 Pro and began rolling it out across developer, enterprise, and consumer surfaces. The post reports a verified ARC-AGI-2 score of 77.1% and lists immediate access via Gemini API, Gemini CLI, Vertex AI, Gemini app, and NotebookLM.