Mistral releases Large 4 for cybersecurity use

AI · First reported 29 Oct, 14:03 · Updated 15:15

Image: Wikimedia Commons · NASA Johnson Space Center · Public domain

Mistral says its publicly released Large 4 model is designed to help organisations find and fix cybersecurity vulnerabilities. The downloadable model can run on users’ own systems; a separate technical report describes its architecture, evaluations and licensing terms.

Mistral says Large 4 is designed to help organisations identify and fix cybersecurity vulnerabilities, and that the public release upgrades a version tested with government and corporate users.
The model can be downloaded and run independently on an organisation’s own systems, reducing reliance on a third-party provider; Mistral says this also means access cannot be contained once released.
A separate technical account reports a one-million-token context window and a modified MIT license that requires a separate commercial license for Model-as-a-Service offerings.

Summary written by Lux AI from 2 articles

Key facts

Release: Mistral publicly released Large 4 on Thursday, October 29, according to the report.
Stated purpose: Help organisations identify and fix cybersecurity vulnerabilities.
Deployment: The model can be downloaded and run independently on users’ systems.
Context window: Up to 1 million tokens, subject to serving memory and runtime limits.
License condition: Model-as-a-Service businesses must obtain a separate commercial license from Mistral.