Maple

Is Mistral open source?

Mistral AI released Mistral 7B under the Apache License 2.0, which is on the Open Source Initiative’s approved list, and Codestral under its Mistral AI Non-Production License, which is not.

2 entries Versions 2 First entry Sep 2023 Latest Checked

Versions and licenses

Newest first · wording from each license file
VersionDateLicenseCommercial use, in the license’s wordsWeightsSourceChecked
Codestral Mistral AI Non-Production License You shall not supply the Mistral Models or Derivatives in the course of a commercial activity, whether in return for payment or free of charge, in any medium or form, including but not limited to through a hosted or managed service (e.g. SaaS, cloud instances, etc.), or behind a software layer. Downloadable License file (opens in a new tab) · Entry
Mistral 7B Apache License, Version 2.0 each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form. Downloadable License file (opens in a new tab) · Entry

“Downloadable” means the weights are published for download under the license named in the row. It is not a statement about that license’s terms.

Dated entries

2024

1 entry

Release: Codestral under the Mistral AI Non-Production License for non-production use, with Apache 2.0 releases continuing.

Mistral AI news, MNPL (primary source, opens in a new tab)

Details
On open weights Codestral is released today under this new license

The May 29, 2024 post says: “This license allows developers to use our technology for non-commercial purposes and to support research work.” It also says: “Note that we will keep releasing models and code under Apache 2.0 as we progressively consolidate two families of products released under Apache 2.0 and the MNPL.” The license limits use to “testing, research, Personal, or evaluation purposes in Non-Production Environments,” and its definition of the model includes “instructed checkpoints, parameters.”

License
Mistral AI Non-Production License (opens in a new tab)
Commercial use
You shall not supply the Mistral Models or Derivatives in the course of a commercial activity, whether in return for payment or free of charge, in any medium or form, including but not limited to through a hosted or managed service (e.g. SaaS, cloud instances, etc.), or behind a software layer.
Weights
Downloadable
Previous license
Apache License, Version 2.0 (opens in a new tab)
Source type
Company page
Checked
Entry
#2024-05-29-mistral-mnpl

2023

1 entry

Release: Mistral 7B under the Apache 2.0 license, which the post says can be used without restrictions.

Mistral AI news, Announcing Mistral 7B (primary source, opens in a new tab)

Details
On open weights We're releasing Mistral 7B under the Apache 2.0 license, it can be used without restrictions.

The September 27, 2023 post is the release notice and includes a download link. The Hugging Face repo mistralai/Mistral-7B-v0.1 lists license apache-2.0 and weight files, and it has no LICENSE file among its siblings.

License
Apache License, Version 2.0 (opens in a new tab)
Commercial use
each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form.
Weights
Downloadable
Previous license
None recorded
Source type
Company page
Checked
Entry
#2023-09-27-mistral-7b

All entries: 2 shown

Free to use

Cite this data

Free to quote, chart and republish under CC BY 4.0. All we ask is a credit with a link back. Sources keep their own terms. CC BY 4.0

Source: Maple, Open-weight AI policy: Mistral (maple.ai/open-weights/mistral), data through 2024-05-29.

Latest entry · Checked

Download the data: data.json · data.csv · How entries are chosen

Maple runs open-weight models inside a hardware-isolated enclave. Maple cannot see your chats. How it works

Every entry cites one primary source, quotes the document’s own words about open weights, and is checked against that source before it publishes. Corrections keep the old wording visible.
Methodology · Corrections · Suggest an entry or a correction

RSS · JSON · CSV
Also kept by Maple: the privacy timeline