Zo | zo.me

This week brought massive advancements in AI - New state-of-the-art models(one called manus that beats deepseek), MCPs changing the multi-agent game, OpenAI's leaked pricing of their upcoming operator models etc.

Here’s a recap of everything you need to know!(Bookmark for later)👇

Anthropic shook up the AI dev world with Claude 3.7 Sonnet—arguably the best coding AI yet.

It can build functional apps in one go, making it feel like a 10x engineer with impeccable taste.

Introducing Claude 3.7 Sonnet: our most intelligent model to date. It's a hybrid reasoning model, producing near-instant responses or extended, step-by-step thinking.

One model, two ways to think.

We’re also releasing an agentic coding tool: Claude Code. https://t.co/jt7qQmFWuC

Anthropic

Zo is the OS for Human to Agent coordination

Zo is a group chat app and protocol powered by AI

Anthropic shook up the AI dev world with Claude 3.7 Sonnet—arguably the best coding AI yet.

It can build functional apps in one go, making it feel like a 10x engineer with impeccable taste.
https://x.com/AnthropicAI/status/1894092430560965029

Meanwhile, OpenAI dropped GPT-4.5, their largest model to date.

It's powerful, witty, and, well... expensive. But Pro users can access it now.

Today we’re releasing a research preview of GPT-4.5—our largest and best model for chat yet.

Rolling out now to all ChatGPT Pro users, followed by Plus and Team users next week, then Enterprise and Edu users the following week. https://t.co/br5win5OEB

OpenAI

Meanwhile, OpenAI dropped GPT-4.5, their largest model to date.

It's powerful, witty, and, well... expensive. But Pro users can access it now.
https://x.com/OpenAI/status/1895219591070261266

ElevenLabs unveiled Scribe, an AI-powered speech-to-text model transcribing 99 languages with near-perfect accuracy.

It even outperforms Whisper v3 and Gemini 2.0 Flash in real-world tests.

Introducing Scribe — the most accurate Speech to Text model.

It has the highest accuracy on benchmarks, outperforming previous state-of-the-art models such as Gemini 2.0 and OpenAI Whisper v3.

It’s now the leading model for English, Spanish, Italian, and many more. With support… https://t.co/A6TzLzFEUL

ElevenLabs

ElevenLabs unveiled Scribe, an AI-powered speech-to-text model transcribing 99 languages with near-perfect accuracy.

It even outperforms Whisper v3 and Gemini 2.0 Flash in real-world tests.https://x.com/elevenlabsio/status/1894821477230485570

Google is making moves, too—Gemini Code Assist now has a free version!

Developers get 180,000 code completions/month, 90x more than Copilot’s free tier, plus deep IDE integration.

We’re launching a free version of Gemini Code Assist globally to help you build faster. It comes with:
🛠️ 180K code completions per month
🌐 Support for all programming languages in the public domain
💡 128K token context window

Get started → https://t.co/w12E5VDPI3 https://t.co/UJv98rre8X

Google DeepMind

Google is making moves, too—Gemini Code Assist now has a free version!

Developers get 180,000 code completions/month, 90x more than Copilot’s free tier, plus deep IDE integration.https://x.com/GoogleDeepMind/status/1894349711160578093

Speed is the name of the game—Inception Labs' Mercury is an ultra-fast LLM, generating text 10x faster than competitors like 4o-mini, Haiku 3.5, and Gemini Flash.

We are excited to introduce Mercury, the first commercial-grade diffusion large language model (dLLM)! dLLMs push the frontier of intelligence and speed with parallel, coarse-to-fine text generation. https://t.co/HfjDdoSvIC

Inception Labs

Speed is the name of the game—Inception Labs' Mercury is an ultra-fast LLM, generating text 10x faster than competitors like 4o-mini, Haiku 3.5, and Gemini Flash.https://x.com/InceptionAILabs/status/1894847919624462794

And in AI video, Alibaba’s Wan2.1 is making waves.

It outperforms Sora in motion realism and physics simulations—while generating 2.5x faster. Oh, and it’s open-source.

I can't believe this is open source.

Wan 2.1 has great prompt coherence and the realism is really impressive for an open source model.

This is image to video by the way. https://t.co/pZSn9Tm4de

Halim Alrasihi

And in AI video, Alibaba’s Wan2.1 is making waves.

It outperforms Sora in motion realism and physics simulations—while generating 2.5x faster. Oh, and it’s open-source.
https://x.com/HalimAlrasihi/status/1895089245301113272

Amazon introduced Alexa+, their most conversational and personalized AI assistant yet.

It dynamically switches between models like Nova and Claude, handling everything from bookings to purchases.

Free for Prime members!

The latest evolution of generative AI is here. Meet Alexa+, our smartest, most conversational, and personalized AI assistant to date.
 
Alexa+ is designed to leverage state-of-the-art architecture that connects LLMs, agentic capabilities, third-party services, and more to your… https://t.co/yJ9IoPeVj0

Amazon

AI design tools have leveled up—Ideogram 2a now generates high-quality images in just 10 seconds, at half the cost.

It’s live now across web, API, and partner platforms.

AI design tools have leveled up—Ideogram 2a now generates high-quality images in just 10 seconds, at half the cost.

It’s live now across web, API, and partner platforms.
https://x.com/0xmetaschool/status/1896112516130095553

Mistral AI launched Mistral OCR, an API for document understanding.

Extracts text from PDFs and images with high accuracy—perfect for RAG pipelines—at just $1 per 1000 pages.

Mistral AI launched Mistral OCR, an API for document understanding.

Extracts text from PDFs and images with high accuracy—perfect for RAG pipelines—at just $1 per 1000 pages.
https://x.com/sophiamyang/status/1897713370029068381

Google is testing AI Mode in Search, using Gemini 2.0 to provide more advanced, contextual responses.

This marks a huge shift in how we’ll interact with search engines.

Google is testing AI Mode in Search, using Gemini 2.0 to provide more advanced, contextual responses.

This marks a huge shift in how we’ll interact with search engines.https://x.com/Google/status/1897332929136877854

Dev tools are also evolving—Codeium’s Windsurf Editor now features live previews, so devs can see website or app updates in real-time.

Dev tools are also evolving—Codeium’s Windsurf Editor now features live previews, so devs can see website or app updates in real-time.https://x.com/windsurf_ai/status/1897378545799979238

OpenAI quietly upgraded ChatGPT’s macOS app, adding direct code editing within IDEs.

Currently for Plus, Pro, and Team users—expanding to Free and Enterprise soon.

OpenAI quietly upgraded ChatGPT’s macOS app, adding direct code editing within IDEs.

Currently for Plus, Pro, and Team users—expanding to Free and Enterprise soon.https://x.com/OpenAIDevs/status/1897700857833193955

Microsoft entered AI-powered healthcare with Dragon Copilot, an assistant for doctors.

It manages voice dictation and automates clinical documentation, saving hours of admin work.

Microsoft entered AI-powered healthcare with Dragon Copilot, an assistant for doctors.

It manages voice dictation and automates clinical documentation, saving hours of admin work.https://x.com/satyanadella/status/1896581949298536757

Tencent’s HunyuanVideo I2V Model extends text-to-video AI to image-to-video generation, with open-source LoRA training for custom effects.

Tencent’s HunyuanVideo I2V Model extends text-to-video AI to image-to-video generation, with open-source LoRA training for custom effects.https://x.com/TXhunyuan/status/1897558826519556325

Voice AI just got an upgrade—Sesame’s hyper-realistic AI voices, Maya and Miles, bring expressive speech synthesis to a new level.

Voice AI just got an upgrade—Sesame’s hyper-realistic AI voices, Maya and Miles, bring expressive speech synthesis to a new level.https://x.com/sesame/status/1895159087010324615

Alibaba’s QwQ-32B: their latest open-source model, optimized for advanced reasoning and rivaling OpenAI’s o1 and DeepSeek-R1.

Alibaba’s QwQ-32B: their latest open-source model, optimized for advanced reasoning and rivaling OpenAI’s o1 and DeepSeek-R1.https://x.com/Alibaba_Qwen/status/1897361654763151544

Creativity meets AI animation—PikaFrames from @pika_labs takes keyframe interpolation to the next level.

The community is already creating stunning results. Try it on their iOS app or website.

Creativity meets AI animation—PikaFrames from @pika_labs takes keyframe interpolation to the next level.

The community is already creating stunning results. Try it on their iOS app or website.https://x.com/pika_labs/status/1895228343496372361

Twitter is flooded with tweets about "MCP," so we've curated the best tweets, demos, and charts to get you up to speed on everything MCP—from A to Z.

Discover why MCP is a step-function leap in AI—all in just 5 minutes.🧵👇

(Or bookmark for later)

Twitter is buzzing about MCP—but what is it?

@zoapp curated the best tweets, demos, and insights to get you up to speed on everything MCP, from A to Z.

Twitter is buzzing about MCP—but what is it?

@zoapp curated the best tweets, demos, and insights to get you up to speed on everything MCP, from A to Z.https://warpcast.com/zoapp/0x42658089

That’s a wrap on this week’s AI agent breakthroughs!

Which release are you most excited about?

And follow us for more such content!

Amazon introduced Alexa+, their most conversational and personalized AI assistant yet.

It dynamically switches between models like Nova and Claude, handling everything from bookings to purchases.

Free for Prime members!
https://x.com/amazon/status/1894796967894479141