Mojo 1.0 Beta

Mojo released its 1.0.0b1 beta. The current stable version is the May 7 release, and the latest nightly is available as of May 9.

Mojo is a statically typed language designed to write like Python and run like C++. It allows writing high-performance kernels in the same language across CPU, GPU, and ASIC, and natively supports Python interoperability, letting developers gradually migrate only the bottlenecks of existing code. Mojo code can be imported into Python, and Python libraries can also be imported into Mojo code.

The core features are compile-time metaprogramming and reflection. It highlights hardware-specific optimization, SIMD vectorization, conditional compilation, and zero-cost abstractions, presenting examples such as a GPU vector_add kernel and an __eq__ implementation using struct field reflection.

The roadmap is divided into Phase 0-3.

• Phase 0: Language foundations such as parser, memory types, functions, and structs
• Phase 1: High-performance kernels for CPU/GPU/ASIC and Python extensions
• Phase 2: Systems applications and guaranteed memory safety
• Phase 3: Dynamic object-oriented features such as classes and inheritance

The stdlib is open-sourced on GitHub, and the compiler is planned to be open-sourced in 2026. Getting-started resources such as Install, Quickstart, Roadmap, GitHub, and llms.txt are also provided.