Tokamax
Tokamax is a library of custom accelerator kernels. It provides highly optimized implementations of key operators, including mixture-of-experts building blocks like grouped matrix multiplication and token gathers/scatters, attention variants like paged attention, splash attention, gated deltanet, Kimi delta attention and multiheaded latent attention, and TopK computation. The operator portfolio will evolve over time to support the latest OSS models running on the latest hardware accelerators, including Google TPUs and NVIDIA GPUs. The library also provides foundational infrastructure to simplify kernel authoring, maintenance, and deployment; for example, features like autotuning, low-overhead benchmarking, and so on.
Tokamax kernels are typically built on top of JAX and Pallas, but also expose native PyTorch APIs on TPUs for developer convenience.
Documentation
Detailed documentation can be found at openxla.org/tokamax, including a list of supported ops on different hardware accelerators, benchmarking infrastructure, and autotuning.
Getting Started
Installation
The latest Tokamax PyPI release:
pip install -U tokamax
The latest bleeding edge version from Github, with no stability guarantees:
pip install git+https://github.com/openxla/tokamax.git
Usage Examples
Please see the Colab notebooks here for usage examples.
Contributing
Please see documentation on how to contribute and our code style guide.
1 maintainer
Metadata
1 maintainer
Release files for tokamax 0.0.17
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| tokamax-0.0.17.tar.gz | 2.1 MB | With provenance | Details for tokamax-0.0.17.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| tokamax-0.0.17-py3-none-any.whl | Python 3 | none | any | Details for tokamax-0.0.17-py3-none-any.whl |
Total release size: 4.9 MB
Release files / tokamax-0.0.17.tar.gz
| Download URL | tokamax-0.0.17.tar.gz |
|---|---|
| Size | 2.1 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
1d20c3d6c5cae4ddf2f09245adf8b35fc5b3c26997b34a4e4b3322d597aa6271
|
|
BLAKE2b-256 checksum How to use checksums |
5500f76565bbf0eae1a88c77ddc23f14dda144e886947bc92a2e6c813aeed5b0
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.7
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 20, 2026.
Transparency logRelease files / tokamax-0.0.17-py3-none-any.whl
| Download URL | tokamax-0.0.17-py3-none-any.whl |
|---|---|
| Size | 2.8 MB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
232a1be65c52b3c436f3836f96889a521c77199df91be8d40a1af63016f0c2d6
|
|
BLAKE2b-256 checksum How to use checksums |
041ca9df63164baf14f4ea4d55695da93630ff86140ad6c6bd2611bcd6b54289
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.7
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 20, 2026.
Transparency logSecurity for tokamax 0.0.17
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
All 2 files in this release have been uploaded with provenance metadata.
Attestation for tokamax-0.0.17.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 20, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.
