Modular

AI coding agents are only as good as the foundation they build on. Our latest 26.2 release ships Mojo 🔥 coding agent skills, purpose-built for writing and porting GPU kernels. Point Claude or Cursor at a CUDA kernel, get idiomatic Mojo back. Also in this release: FLUX.2 image gen at 3-4x PyTorch Diffusers speed with a 4.1x cost savings on B200. See everything in 26.2 ↓ modular.com/blog/modular-2…

Oh, and we're hiring! See open roles: #open-roles" target="_blank" rel="nofollow noopener">modular.com/company/career…

Modular is officially in Edinburgh. We're at the Bayes Centre, where world-leading data science and AI teams work alongside businesses to turn research into real solutions. A fitting place to build the next layer of AI infrastructure.

130 lines instead of 870. That's the difference between our conv2d implementation on Blackwell and CUTLASS's. We broke kernels into three swappable pieces: one for moving data, one for coordinating the pipeline, one for compute. When you need a new kernel, you only change the piece that actually needs to change. Part 3 of our Structured Mojo Kernels series walks through the details: modular.com/blog/structure…

Our March community meeting recording is now live! Recap includes: community member Mohamed Mabrouk on BlazeSeq, a zero-copy GPU-friendly FASTQ parser written in Mojo 🔥 hitting 5+ GB/s, variadic metaprogramming in Mojo and 26.2 release overview. youtube.com/watch?v=Sur7e3…

@swyx 🙇

@Modular we are not worthy

2 days ago we shipped image generation in <1s 🔥 Today, we make that <300ms 🤯 NVIDIA + AMD⚡️ Full demo below ⬇️

Reach out to us on modular.com if you get this 🚀

Full demo ⬇️

"We're in a tidal wave of new chips. And we, as software engineers, need to be able to program all of them." Our CEO and co-founder @clattner_llvm joined @alexeheath from Sources live at GTC to talk hardware portability, the expanding chip landscape, and what Modular is shipping to make it possible. sources.news/p/trying-to-br…

Here they are: github.com/modular/skills

That moment you use our new AI skills 💻🧑‍💻⬇️

@thaoshibe @MuCai7 NanoBanana 🍌 is so good, it turned into NanoBanna🥷

@MuCai7 @Modular i want to react "haha" but unfortunately there is no "haha" function on x hahaa

Generate images in less than 1 second. 99% cheaper than NanoBanna. 🚀 😱 Our latest 26.2 release ships FLUX.2 image generation with a 4.1x speedup over torch.compile on NVIDIA Blackwell - translating to a 5.5x TCO advantage with AMD MI355X. Read more ⬇️

@morqon Use the Mojo, Luke. You know it to be true. 🔥

might write my first kernel

Cats on MAX ... because of course! 😏

Image gen in <1s 🤯 You asked for a demo - here's @iamtimdavis showing Flux 2 Dev on the Modular stack. Reach out to us if you're interested! 🚀 But thats not all - video generation is coming ⬇️

@jquave @clattner_llvm Fixed 🔥 Come talk to us!

@Modular @clattner_llvm found an issue with your platform. show stopper in fact. it’s pictured here

@Zyyon_ Mojo 🔥 is what powers it all :) Checkout our latest blog for more details ➡️ modular.com/blog/modular-2…

Hold on What happened to Mojo?

Full Blog: modular.com/blog/modular-2…

FLUX.2 demo: youtube.com/watch?v=GzLYJM…