OpenAI GPT-OSS-120B is live on Cerebras 3,000 tokens/s - fastest OpenAI model on record 1 second reasoning time 131K context Go build something fantastic. inference.cerebras.ai

This is a watershed moment - OpenAI's open source model - available day0 on the Vercel AI Gateway. This is the open model enterprises have been waiting for - 20B and 120B parameters, open weights powered by OpenAI!

Try both OpenAI open source models on Vercel AI Playground here: ai-sdk.dev/playground/gro… and compare to any of the other 100+ models in our AI Gateway

You can now access Claude Opus 4.1 through Vercel AI Gateway. Set the AI SDK v5 𝚖𝚘𝚍𝚎𝚕 to '𝚊𝚗𝚝𝚑𝚛𝚘𝚙𝚒𝚌/𝚌𝚕𝚊𝚞𝚍𝚎-𝟺.𝟷-𝚘𝚙𝚞𝚜' to start building. Inference powered by @anthropicai and Amazon Web Services Bedrock with no accounts required. vercel.com/changelog/clau…

So much fun presenting our OpenBlock Agent Arena at @Vercel HQ! Also got to see an insanely cool v0 launch that’s coming very soon 🔥 s/o alli Guillermo Rauch for hosting!

Busy day for new model releases: Claude Opus 4.1, the many gpt-oss-* variants, Qwen Image! Try out and compare these new models and more for yourself for free. They were all available on yupp.ai within minutes of their announcements.🔥

I love that Vercel have amazing support for Hono

Super useful capability for all Vercel customers!

Earlier today we released the Vercel MCP server, a read only remote MCP server that you can use from Claude Code, Copilot ...(more added daily). Initial use cases include debugging deployments and providing LLM clients the latest documentation.

From on-demand software to software on demand. Vibe coding with AI FTW

Can't believe I'm seeing Guillermo Rauch live vibe coding a micro-blogging app with v0 powered by subscriptions on Porto

If you’re using AI SDK and AI Gateway, you replace one string to ‘𝚘𝚙𝚎𝚗𝚊𝚒/𝚐𝚙𝚝-𝟻’ and you’re live. No signup required. As an aside, I don’t think I’ve ever seen an industry that moves this fast, with the switching costs this low. The best intelligence wins.

Today we’re open sourcing a “vibe coding agent” powered by GPT-5. It’s like v0, but agnostic to framework, language, runtime. It can vibe code htmx and Haskell if you want. Built on AI SDK, Sandbox and AI Gateway. If you want to add codegen to your platform or build your own

Vibe coding platform is out We had so much fun building this with Javi Velasco, so happy to open source it Powered by Sandbox for the untrusted code, Fluid to run everything efficiently, and AI gateway for multiple models Built with Next.js and AI SDK oss-vibe-coding-platform.vercel.app

I built the first iteration of this project during an internal hackathon a couple of weeks ago. Guillermo Rauch thought it was a pretty cool showcase for some of our tech—like the AI Gateway, AI SDK, Sandbox, Fluid Functions. So we started iterating on that early version, with one of

Things are changing continuously on our newly launched leaderboard. Gpt-5 currently having a peak moment. AI Gateway share.google/bEhJ3X6x3IdNRG…

Just updated the Vercel MCP server to add Cursor clients. We will keep adding more clients. Continues to be a read-only, remote MCP server w/ built-in oauth. Give your AI client up-to-date Vercel docs and monitoring for Vercel deployments. vercel.com/changelog/curs…

v0.dev is now v0.app, the AI builder for everyone. The new v0 uses agentic AI to plan, research, build, and debug. It works in context, across multiple steps, and adapts as you go. If you have an idea, v0 can build it.

Get your free v0 credits by being one of the first users daily this week. And try out the new agent-mode