Latest Generative AI stories and analysis from TECH CURRENT.
On October 1, 2026, the Strands Agents team released Strands Decider 2B, an open-source two-billion-parameter model designed to choose among predefined options rather than generate free-form text.
Read article →On September 3, 2026, Google introduced Gemini 3.8 Flash and Gemini 3.8 Flash Cyber. The standard model focuses on reasoning, coding and multi-step agent tasks.
Read article →MentalHealthBench uses expert-written criteria from more than 80 licensed professionals in 22 countries to evaluate AI responses.
Read article →On September 9, 2026, OpenAI called for mandatory national AI safety requirements tied to model capabilities. The company argued that industry self-governance alone is insufficient as advanced systems
Read article →Google's Gemini 4 Argon targets long-horizon reasoning in software engineering, enterprise workflows and defensive cybersecurity, with a phased rollout.
Read article →A look at Google's EmbeddingGemma 2, a compact model that embeds text, code, images, video and audio for local search and retrieval.
Read article →OpenAI introduced Intelligent UI for GPT-6 on October 7, 2026, bringing interactive charts, controls and tools into ChatGPT conversations.
Read article →OpenAI has detailed textGrain watermarking for AI-generated text in response to EU rules, while acknowledging detection failures after editing and other limitations.
Read article →Google introduced Gemini 3.5 Flash at I/O 2026 and made it generally available through developer tools including the Gemini API.
Read article →Gemini Omni was introduced as a multimodal model that can create and edit content from different kinds of inputs, beginning with video.
Read article →