Thinking Machines has previewed Interaction Models for near-real-time voice, video and text AI, built to listen, speak, see and use tools simultaneously.
Alibaba is preparing to integrate Qwen with Taobao and Tmall, giving the AI app access to more than 4 billion products as Taobao Shopping moves toward agentic commerce.
Claude memory can reduce repetitive setup prompts by carrying useful context across conversations, including preferences, project details, tone, and workflow rules.
OpenAI has launched three new voice API models for live audio agents, real-time translation and streaming transcription, with GPT-Realtime-2 adding larger context, adjustable reasoning effort, tool transparency and stronger recovery behaviour.
OpenAI has introduced GPT-5.5-Cyber, a limited-preview model for vetted cybersecurity professionals, as Anthropic investigates reported unauthorized access to Claude Mythos.
Anthropic introduced dreaming for Claude Managed Agents, a research preview feature that lets agents review past sessions, extract patterns, update context files, and create reusable playbooks for future work.
Microsoft, Google DeepMind and xAI have agreed to give U.S. officials early access to new AI models for pre-release AI model testing focused on national security risks.
OpenAI is making GPT-5.5 Instant the default ChatGPT model, reporting fewer hallucinated claims, stronger benchmark results, low latency and new controls for personalized memory.
IBM released Granite 4.1, its broadest Granite model family to date, with new language, speech, vision, embedding and Guardian models for enterprise AI workflows.
