Anthropic released Claude Opus 5 with a 1M-token context window and near-Fable-5 performance at the same price as Opus 4.8. Google announced that Gemini 4 pre-training has already begun β described as the most ambitious training run in Google history. And Claude Science, Anthropic's AI workbench for scientists connected to 60+ research databases, is quietly redefining what it means to do research with AI.
NotebookLM rebranded as Gemini Notebook and gained native code execution. Claude voice mode now supports Opus and Sonnet models and connects to Gmail, Slack, and more. Google launched Gemini 3.6 Flash β 17% fewer output tokens, lower price, higher benchmarks.
Claude Code v2.1.217 raises the default nested subagent depth to 3, enabling true multi-tier agentic pipelines. Google Vids integrates Gemini Omni to generate personal AI avatar videos from a selfie and edit existing videos with plain text commands. Suno completely rebuilt its Lyrics Editor, adding a Lyricist profile that remembers your writing style and natural-language editing.
Google released Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber simultaneously β the flagship 3.6 Flash cuts token usage by up to 17% while improving coding and knowledge-work performance. Anthropic added a Reflect dashboard to Claude, letting users visualize their own AI usage patterns and topics. Notion expanded its Agents feature to iOS and introduced calendar management tools so agents can schedule meetings, send invites, and find open time slots.
Claude Code's default model has switched to Sonnet 5, bringing a 1M-token context window out of the box. Gemini Deep Research now accepts your own PDFs and images as research sources, and Canvas can transform the resulting reports into interactive quizzes and infographics. Vibe coding has become the standard workflow for anyone who wants to build apps with Claude Code β no programming background required.