Modal

Modal provides a high-performance cloud platform for developers, allowing you to bring your own code for CPU, GPU, and data-intensive compute tasks without worrying about infrastructure. The platform is designed specifically for AI, data, and machine learning (ML) teams, offering the ability to run generative AI models, large-scale batch jobs, and job queues with efficiency and speed. Modal's serverless environment is crafted with a container system built in Rust, ensuring fast cold-start times, scalability to hundreds of GPUs, and a cost-effective model that charges only for actual compute usage by the CPU cycle. Features like effortless deployment, comprehensive environments, built-in storage, and job scheduling enable you to focus entirely on your code, while security and observability are handled by the platform, promoting seamless and secure development experiences.

Key Features

• Custom Container System: Built from scratch in Rust for fast cold-start times and scalability.

• Serverless Platform: Specifically tailored for AI and data teams with no infrastructure management required.

• Flexible Environments: Express container images and hardware specifications entirely in code without Dockerfiles or YAML.

• Comprehensive Developer Experience: Provision storage, schedule jobs, serve HTTPS endpoints, and secure workloads easily.

• Cost-Effective: Pay only for actual compute usage and get $30 of compute credits each month.

FAQ

What is Modal?

Modal is a serverless platform that allows developers to run their code on CPU or GPU for AI, data, and machine learning tasks without managing infrastructure.

What are some key features of Modal?

Modal features include a custom container system, a serverless environment, comprehensive developer tools, and cost-effective pricing with $30 free compute each month.

How does Modal charge for compute resources?

Modal charges for compute resources by the CPU cycle and offers competitive rates for GPU usage with different Nvidia models, including H100, A100, A10G, L4, and T4.

What kind of compute tasks can I run on Modal?

You can run a variety of compute tasks on Modal including generative AI models, large-scale batch jobs, and job queues.

Are there any special offers for startups or academic researchers?

Startups and academic researchers are eligible to apply for up to $25k in free compute credits, which can be used towards GPU compute and accessing in-demand GPU types.