Google released Gemini 3.8 Live and a new Extended Thinking variant, expanding real-time AI capabilities with enhanced reasoning for complex problem-solving tasks.
Google announced Gemini 3.8 Live, an updated version of its real-time AI model, alongside Gemini 3.8 Live Extended Thinking. The Extended Thinking variant introduces deeper reasoning capabilities designed for complex analytical work.
Gemini 3.8 Live maintains the low-latency, streaming response functionality of its predecessor, enabling near-instantaneous interactions. The Extended Thinking version trades some speed for improved reasoning depth, allowing the model to work through multistep problems with more thorough analysis before providing responses.
The release targets developers and enterprises seeking different performance profiles. Gemini 3.8 Live suits applications requiring quick interactions—conversational AI, real-time translation, and live customer support. Extended Thinking addresses scenarios demanding higher accuracy over speed, such as data analysis, research assistance, and technical problem-solving.
Google positioned these models as part of its broader Gemini family, which includes various sizes and specializations. The company emphasized improved performance benchmarks and expanded context windows across both variants.
The announcement generated significant developer interest. Hacker News discussion drew 122 comments with 184 upvotes, focusing on practical applications, API pricing, and comparisons with competing models like OpenAI's offerings. Developers discussed trade-offs between latency and reasoning quality, with Extended Thinking's performance on standardized benchmarks emerging as a key topic.
Both models are available through Google's AI Studio and Vertex AI platform. Pricing follows Google's existing tiered structure based on input/output tokens, with Extended Thinking commanding higher rates due to increased computational requirements.
The release reflects ongoing competition in large language models, where companies balance speed, accuracy, and cost. Google's dual-variant approach lets customers optimize for their specific use cases rather than choosing a single fixed model.
Treasury Secretary Scott Bessent told House lawmakers that AI developers should face full legal liability for their systems and that the US should prioritize open-source model development.
Typesafe has introduced System One Models and Jev, new tools designed to improve type safety and developer experience. The announcement generated significant discussion with 375 points and 145 comments on Hacker News.
BlackRock CEO Larry Fink warned that delays in artificial intelligence infrastructure buildout are restricting who can access the technology. Fink made the comments at the Canada Investment Summit on Tuesday.
Anthropic and OpenAI's coordination with the US government on AI safeguards could create barriers for smaller competitors, according to startup executives and industry observers.