USASI
SoftwareResearch stack

MaxText

Project record

Maintained by Google (AI Hypercomputer GitHub organization)245

MaxText is an open-source LLM library and reference implementation written in Python and JAX for training on Google Cloud TPUs and GPUs. It provides implementations of Gemma, Llama, DeepSeek, Qwen, Mistral, and other open model architectures, with support for pre-training and for post-training methods such as supervised fine-tuning and reinforcement learning.12

Last reviewedEntry updated Documented release Unknown

Availability and license

Overall availability

Public

Documented as available to the general public. Access conditions and license terms may still apply.13

Availability is separate from permission: read the license before using or redistributing.

The Apache 2.0 license covers the MaxText repository. This record does not assess the licenses of the third-party model checkpoints that MaxText can load or train.32

Public materials checklist

Items for a research stack under USASI rubric v0.1. Unknown means unassessed or insufficient evidence.
Public materials checklist for MaxText
ItemStatusNotes and evidence
Source codeIs the source code publicly readable?PublicPublic GitHub repository under the Apache License 2.0.13
DocumentationIs user documentation published?PublicDocumentation is published on Read the Docs.62
Training codeDoes it include code for training models?PublicIncludes code for pre-training and for post-training (supervised fine-tuning and reinforcement learning).2
Data informationAre the data it expects or ships with documented?PublicThe documentation describes the supported data input pipelines (Grain, Hugging Face, and TFDS) and the file formats each accepts.7
ReproducibilityAre instructions for reproducing reported results published?UnknownGetting-started and tutorial guides are published; whether they reproduce specific reported performance results was not assessed.2

What it is useful for

The README presents MaxText as a starting point to fork for pre-training experiments on TPU or GPU and as a framework for post-training existing models.2

Run and use notes

Documented facts only. No hardware or performance claims are made without a cited source and stated assumptions.
  • The README recommends installing the latest PyPI release or a pre-built container image rather than the main branch, and names Python 3.12 as the primary supported version. PyPI lists version 0.2.4, released August 21, 2026.28

Organization context

U.S. eligibility

Project eligibility rests on documented governing or maintaining entities, not on contributors.

Eligible · basis: U.S.-governed project

MaxText's source files carry a Google LLC copyright notice, its contribution guide requires Google's Contributor License Agreement and follows Google's Open Source Community Guidelines, and it is hosted in the AI Hypercomputer GitHub organization, which describes itself as covering Google Cloud AI Hypercomputer. Google LLC is a Delaware subsidiary of Alphabet Inc., whose principal executive offices are in Mountain View, California, per its fiscal 2025 Form 10-K.245910

Assessed Sep 29, 2026

Sources

  1. 1.
    AI-Hypercomputer/maxtext (GitHub) (external site: github.com)

    Google (AI Hypercomputer) · Repository · accessed Sep 29, 2026

  2. 2.
    MaxText README.md (external site: raw.githubusercontent.com)

    Google (AI Hypercomputer) · Repository · accessed Sep 29, 2026

  3. 3.
    MaxText LICENSE (Apache License 2.0) (external site: raw.githubusercontent.com)

    Google (AI Hypercomputer) · License · accessed Sep 29, 2026

  4. 4.
    MaxText CONTRIBUTING.md (external site: raw.githubusercontent.com)

    Google (AI Hypercomputer) · Repository · accessed Sep 29, 2026

  5. 5.
  6. 6.
    MaxText documentation (external site: maxtext.readthedocs.io)

    Google (AI Hypercomputer) · Documentation · accessed Sep 29, 2026

  7. 7.
    MaxText data input pipelines (external site: maxtext.readthedocs.io)

    Google (AI Hypercomputer) · Documentation · accessed Sep 29, 2026

  8. 8.
    maxtext on PyPI (external site: pypi.org)

    Python Package Index · Release notes · accessed Sep 29, 2026

  9. 9.
    Alphabet Inc. Form 10-K for the fiscal year ended December 31, 2025 (external site: sec.gov)

    Alphabet Inc. (via U.S. Securities and Exchange Commission EDGAR) · Filing · published Feb 5, 2026 · accessed Sep 29, 2026

  10. 10.
    Exhibit 21.01 — Subsidiaries of the Registrant (Alphabet Inc. Form 10-K, fiscal 2025) (external site: sec.gov)

    Alphabet Inc. (via U.S. Securities and Exchange Commission EDGAR) · Filing · published Feb 5, 2026 · accessed Sep 29, 2026

This listing is not an endorsement, a safety assessment, or a federal approval.

Support Us

Help keep USASI useful.

Optional. No USASI account required. Payment takes place on the linked provider’s website (Buy Me a Coffee).

About supporting this project