🔥 Modular has acquired BentoMl to extend our open source AI stack! Join us on February 17th at 9:30-11:30am PT for an Ask Us Anything with Chris Lattner and BentoML Founder Chaoyu Yang in the Modular forum. We’ll answer questions and share more about our plans! Feel free to share your questions now.
A unified platform for AI development and deployment, including MAX🧑🚀 and Mojo🔥.
The Modular Platform is an open and fully-integrated suite of AI libraries and tools that accelerates model serving and scales GenAI deployments. It abstracts away hardware complexity so you can run the most popular open models with industry-leading GPU and CPU performance without any code changes.
You don't need to clone this repo.
You can install Modular as a pip or conda package and then start an
OpenAI-compatible endpoint with a model of your choice.
To get started with the Modular Platform and serve a model using the MAX framework, see the quickstart guide.
Note
Nightly vs. stable releases
If you cloned the repo and want a stable release, run
git checkout modular/vX.X to match the version.
The main branch tracks nightly builds, while the stable branch matches
the latest released version.
After your model endpoint is up and running, you can start sending the model inference requests using our OpenAI-compatible REST API.
Try running hundreds of other models from our model repository.
The MAX container is our Kubernetes-compatible Docker container for convenient deployment, which uses the MAX framework's built-in inference server. We have separate containers for NVIDIA and AMD GPU environments, and a unified container that works with both.
For example, you can start a container for an NVIDIA GPU with this command:
docker run --gpus=1 \
-v ~/.cache/huggingface:/root/.cache/huggingface \
-p 8000:8000 \
modular/max-nvidia-full:latest \
--model-path google/gemma-3-27b-itFor more information, see our MAX container docs or the Modular Docker Hub repository.
We're constantly open-sourcing more of the Modular Platform and you can find all of it in here. As of May, 2025, this repo includes over 450,000 lines of code from over 6000 contributors, providing developers with production-grade reference implementations and tools to extend the Modular Platform with new algorithms, operations, and hardware targets.
It's quite likely the world's largest repository of open source CPU and GPU kernels!
Highlights include:
- Mojo standard library: /mojo/stdlib
- MAX GPU and CPU kernels: /max/kernels (Mojo kernels)
- MAX inference server: /max/python/max/serve (OpenAI-compatible endpoint)
- MAX model pipelines: /max/python/max/pipelines (Python-based graphs)
- Code examples: /max/examples + /mojo/examples
This repo has two major branches:
-
The
mainbranch, which is in sync with the nightly build and subject to new bugs. Use this branch for contributions, or if you installed the nightly build. -
The
stablebranch, which is in sync with the last stable released version of Mojo. Use the examples in here if you installed the stable build.
We accept contributions to the Mojo standard library, MAX AI kernels, MAX model architectures, code examples, Mojo docs, and more.
First, please read the Contribution Guide, and then refer to the following documentation about how to develop in the repo:
/max/docs: Docs for developers working in the MAX framework codebase./mojo/stdlib/docs: Docs for developers working in the Mojo standard library.
We also welcome your bug reports. If you have a bug, please file an issue here.
[2026/2] We announced that BentoML is joining Modular. We are committed to building in the open and will be extending our support of open source AI with Bento's own open project. Join us for an AMA February 17th to discuss our plans.
[2026/1] Modular Platform 26.1 graduates the MAX Python API out of experimental with PyTorch-like eager mode and model.compile() for production, stabilizes the MAX LLM Book, and expands Apple silicon GPU support. Mojo gains compile-time reflection, linear types, typed errors, and improved error messages as it progresses toward 1.0.
[2025/12] The Path to Mojo 1.0 was officially announced with a planned release in H1 2026 and tons of details on what to expect.
[2025/12] We hosted our Inside the MAX Framework Meetup reintroducing the MAX framework and taking the community through upcoming changes.
[2025/11] Modular Platform 25.7 provides a fully open MAX Python API, expanded hardware support for NVIDIA Grace superchips, improved Mojo GPU programming experience, and much more.
[2025/11] We met with the community at PyTorch 2025 + the LLVM Developers' Meeting to solicit community input into how the Modular platform can reduce fragmentation and provide a unified AI stack.
[2025/09] Modular raises $250M to scale AI's unified compute layer, bringing Modular's total raise to $380M at a $1.6B valuation.
[2025/09] Modular Platform 25.6 delivers a unified compute layer spanning from laptops to datacenter GPUs, with industry-leading throughput on NVIDIA Blackwell (B200) and AMD MI355X.
[2025/08] Modular Platform 25.5 introduces Large Scale Batch Inference through a partnership with SF Compute + open source launch of the MAX Graph API and more.
[2025/08] We hosted our Los Altos Meetup featuring talks from Chris Lattner on democratizing AI compute and Inworld AI on production voice AI.
[2025/06] AMD partnership announced — Modular Platform now generally available across AMD's MI300 and MI325 GPU portfolio.
[2025/06] Modular Hack Weekend brought developers together to build custom kernels, model architectures, and PyTorch custom ops with Mojo and MAX.
[2025/05] Over 100 engineers gathered at AGI House for our first GPU Kernel Hackathon, featuring talks from Modular and Anthropic engineers.
We host regular meetups, hackathons, and community calls. Join us!
| Channel | Link |
|---|---|
| 💬 Discord | discord.gg/modular |
| 💬 Forum | forum.modular.com |
| 📅 Meetup Group | meetup.com/modular-meetup-group |
| 🎥 Community Meetings | Recordings on YouTube |
Upcoming events will be posted on our Meetup page and Discord.
If you'd like to chat with the team and other community members, please send a message to our Discord channel and our forum board.
This repository and its contributions are licensed under the Apache License v2.0 with LLVM Exceptions (see the LLVM License). Modular, MAX and Mojo usage and distribution are licensed under the Modular Community License.
You are entirely responsible for checking and validating the licenses of third parties (i.e. Huggingface) for related software and libraries that are downloaded.