Weekly AI Roundup: Google, Anthropic, OpenAI: New Models, Browser Integration, and Hardware Advances

Kabui, Charles

Download as PDF

Google News

Project Genie Open-World Builder: An AI model that creates interactive, open-world environments where users can control characters and explore generated spaces. Think of it like a video game that builds itself as you play, letting you walk around and interact with things it just “imagined.”, https://blog.google/innovation-and-ai/models-and-research/google-deepmind/project-genie/
Gemini in Chrome Sidebar: Google has integrated Gemini directly into the Chrome browser with “agentic” features, allowing it to reimagining images on a page, fill out complex forms, and draft emails based on what you are viewing. A smart assistant that lives on the side of your internet browser and can help you do chores like filling out boring paperwork or editing pictures you find online, https://blog.google/products-and-platforms/products/chrome/gemini-3-auto-browse/
Google Photos “Me Meme”: A generative AI feature in Google Photos that allows users to place their own faces into popular or custom meme templates. A fun tool that lets you swap your face into funny internet pictures (memes) with just one click, https://support.google.com/photos/thread/401524051/star-in-your-own-memes-introducing-me-meme-in-google-photos?hl=en

Anthropic & Claude Ecosystem

Claude Connectors & Tools: Claude now supports direct integration with apps like Figma, Slack, and Asana via the Model Context Protocol (MCP). Instead of copying and pasting, the AI can “reach” into your other apps (like Slack) to read your messages or help you organize your work tasks, https://www.theverge.com/news/867673/claude-mcp-app-interactive-slack-figma-canva
Official Claude Excel Add-in: A new official tool that brings Claude’s reasoning models (Opus and Sonnet) directly into Microsoft Excel spreadsheets. You can now ask the AI to clean up, analyze, or explain your data directly inside your Excel file without leaving the app, https://support.claude.com/en/articles/12650343-claude-in-excel
OpenClaw Rebrand: The viral open-source assistant formerly known as Claudebot or Moltbot has officially rebranded to OpenClaw following trademark requests from Anthropic. A free, fan-made personal assistant program had to change its name so people wouldn’t confuse it with the official “Claude” company, https://mashable.com/article/clawdbot-changes-name-to-moltbot-openclaw

OpenAI Developments

OpenAI Prism Science Tool: A free AI tool powered by GPT-5.2 designed specifically to help scientists and researchers with technical writing and data analysis. A specialized version of ChatGPT that is “extra smart” at science and helping experts write complicated research papers, https://openai.com/index/introducing-prism/
Retirement of GPT-4o Models: OpenAI is phasing out several older versions of the 4o model in ChatGPT by February 13, 2026, to focus on the newer GPT-5 series. The company is “turning off” older versions of the AI to make room for the newest, most advanced versions, https://yourstory.com/ai-story/openai-retiring-gpt-4o-older-models-chatgpt
Ads in ChatGPT: OpenAI is beginning to roll out advertisements for free users and lower-tier subscribers, reportedly at a premium rate of $60 CPM. If you use the free version of ChatGPT, you will start seeing commercials, similar to how ads appear on YouTube or Facebook, https://www.thekeyword.co/news/openai-sets-early-chatgpt-ad-pricing-at-about-60-cpm

New Models & Hardware

Kimi K2.5 Agentic Model: A new open-source model from Moonshot AI that has topped major benchmarks like “Humanity’s Last Exam” for its ability to navigate the web and use tools. A powerful “free-to-use” AI from China that is currently better at browsing the web and doing tasks than many of the paid US models, https://vertu.com/lifestyle/kimi-k2-5-vs-claude-opus-4-5-why-this-open-source-giant-is-the-new-king-of-agentic-ai/
Qwen3 Max Thinking: Alibaba’s latest high-reasoning model that competes with GPT-5.2 and Claude 4.5 on complex logic tasks. Another very smart AI from Alibaba that specializes in “thinking through” hard math and coding problems, https://artificialanalysis.ai/models/qwen3-max-thinking
Microsoft Maia 200 Chip: Microsoft has introduced its latest custom-built AI chip designed specifically to speed up AI “thinking” (inference) in its data centers. Microsoft made its own “brain chips” for its computers so they can run AI faster and cheaper than using other companies’ parts, https://rcpmag.com/articles/2026/01/29/microsoft-unveils-maia-200.aspx

Video & Multimedia

Luma AI Ray 3.14 Upgrade: An update to Luma’s video model that provides native 1080p generation and is significantly faster and cheaper than previous versions. A tool that makes high-quality videos from your text descriptions just got better, faster, and much more affordable, https://lumalabs.ai/blog/news/ray3_14
Grok Imagine Video API: xAI has released a new API for its Grok model, which has reportedly jumped to the top of the “Video Arena” benchmarks for visual quality. Elon Musk’s AI company released a new tool that creates some of the most realistic-looking videos currently available, https://x.ai/news/grok-imagine-api
Lucy 2 Face-Swapping: A new real-time tool from Decart that allows for highly realistic face-swapping and facial animation for digital avatars. Software that lets you “wear” someone else’s face digitally in real-time, which is popular for online streamers and virtual YouTubers, https://lucy.decart.ai/

Strategic Tech News

Yahoo Scout Search Engine: Yahoo has entered the AI search race with “Scout,” a new engine designed to provide direct answers instead of just a list of links. Yahoo is making a comeback with a search engine that talks back to you and gives you the answer directly, like a digital encyclopedia, https://www.yahooinc.com/press/introducing-yahoo-scout-a-new-ai-answer-engine
Apple Acquires Cue AI: Apple has reportedly spent $2 billion to acquire tech that focuses on silent communication and reading facial expressions. Apple bought a company that helps computers understand what you’re thinking just by looking at your face or mouth, even if you don’t speak out loud, https://www.timesofisrael.com/apple-buys-israeli-startup-reportedly-for-2-billion-that-reads-facial-micro-movements/
NVIDIA Earth-2 Weather Models: NVIDIA has released open-source AI models capable of 15-day precision weather forecasting to help combat climate change impacts. NVIDIA built a “digital twin” of the Earth that uses AI to predict the weather much more accurately than old methods, https://blogs.nvidia.com/blog/nvidia-earth-2-open-models/

Disclaimer: For information only. Accuracy or completeness not guaranteed. Illegal use prohibited. Not professional advice or solicitation. Read more: /terms-of-service

Reuse

GNU GENERAL PUBLIC LICENSE v3.0(View License)

Citation

BibTeX citation:

@misc{kabui2026,
  author = {{Kabui, Charles}},
  title = {Weekly {AI} {Roundup:} {Google,} {Anthropic,} {OpenAI:} {New}
    {Models,} {Browser} {Integration,} and {Hardware} {Advances}},
  date = {2026-02-01},
  url = {https://toknow.ai/posts/weekly-ai-roundup-google-anthropic-openai-new-models-browser-integration-hardware-advances/},
  langid = {en-GB}
}

For attribution, please cite this work as:

Kabui, Charles. 2026. “Weekly AI Roundup: Google, Anthropic, OpenAI: New Models, Browser Integration, and Hardware Advances.” https://toknow.ai/posts/weekly-ai-roundup-google-anthropic-openai-new-models-browser-integration-hardware-advances/.