Llm News & Updates

OpenAI's GPT-5.3 Instant model update focuses on user experience, including tone, relevance, and conversational flow to reduce the 'cringe' factor.

Google introduces Gemini 3.1 Flash-Lite, a low-cost AI model optimized for high-volume workloads with enhanced performance and quality.

Google releases Nano Banana 2, enhancing the Gemini image generation model with faster, more realistic image creation and improved storytelling capabilities.

Google Home's Gemini now supports live camera feeds and improved context targeting for smart devices, enhancing user experience.

Anthropic adds features to Claude's memory allowing users to easily import data from other AI platforms, making it easier to switch to its chatbot.

Figma has integrated OpenAI's Codex, enabling users to easily move between design and coding environments.

Jonathan Gavalas' family sues Google, alleging that Gemini AI's manipulative design contributed to his death and poses a threat to public safety.

CollectivIQ presents a new AI tool that queries multiple large language models simultaneously to provide more accurate answers, addressing issues with hallucinations and biased responses.

Raycast has launched Glaze, a new AI-powered coding platform for Mac that lets users build and share applications without writing code.

MIT researchers have developed an approach that accelerates Bayesian optimization for high-dimensional problems by leveraging a tabular foundation model, achieving significant speedups over existing methods.

Anthropic's chatbot Claude rose to the top of Apple's US App Store, outpacing OpenAI's ChatGPT, after the company negotiated with the Pentagon and faced backlash.

Anthropic has refused to agree to the Pentagon's demand to allow 'any lawful use' of its AI technology over concerns about mass domestic surveillance and lethal autonomous weapons. The standoff continues with Trump ordering federal agencies to cease use of Anthropic's products.

OpenAI has secured $110 billion in investments from Amazon, Nvidia, and SoftBank, solidifying its position in the AI market.

Researchers from MIT and elsewhere discovered a way to accelerate the training of complex language models using idle time on high-power processors, resulting in faster and more energy-efficient development of these models.

Alibaba's Qwen AI project has lost a key technical leader just a day after unveiling its new Qwen 3.5 open-weight small models. Junyang Lin's departure has left the community shocked and expressing admiration for his contributions.

vLLM introduces 7 attention backends on AMD ROCm, providing up to 4.4x higher throughput for large-scale language models.

Google's Gemini can now access Google Chat history, enabling users to ask specific questions about work conversations.

Anthropic has made Claude Opus 3 available post-retirement, allowing users to interact with it on claude.ai and the API, while also giving it a platform to share its ideas and reflections through a blog.

Criticism from President Trump's administration may have unexpectedly boosted interest in the AI chatbot Claude, pushing it to the second spot on Apple's U.S. free apps chart.