Release Notes
Abstract
This technical note provides guidance for installing IBM Guardium Data Protection Small Language Model patch 12.0p85240, including new features and notices associated with the patch.
Content
Patch information
- Patch file name: SqlGuard-12.0p85240_slm_Sep_03_2026
- MD5 checksum: 56bc7f189749299947496703400ef437
Finding the patch
- Select the following options to download this patch on the IBM Fix Central website and click Continue.
- Product selector: IBM Security Guardium
- Installed Version: 12.2
- Platform: All
- On the "Identify fixes" page, select Browse for fixes and click Continue.
- On the "Select fixes" page, select Other Patch. Then, enter the patch information in the Filter fix details field to locate the patch.
For information about Guardium patch types and naming conventions, see the Understanding Guardium patch types and patch names support document.
Prerequisites
- Guardium Data Protection 12.2.4 (GPU 12.0p7240; see release notes)
- The latest Guardium Data Protection health check patch 12.0p9997 (see release notes)
- Guardium Data Protection Gen AI patch 12.0p81240 (see release notes)
Installation
Notes: This patch will not restart any services.
Overview:
- Download the patch and extract the compressed package outside the Guardium system.
- Be sure to check the latest version of these patch release notes online just before you install this patch.
- Pick a "quiet" or low-traffic time to install the patch on the Guardium system.
- Install patches in a top-down manner on all Guardium systems. This patch must be installed starting with the central manager and then installed on the aggregators.
For more information, see Installing patches and Configuring Small Language Model in the Guardium documentation.
New features and enhancements
Small Language Model for AI use cases
Guardium generative AI (gen AI) appliance now includes a pre-packaged Small Language Model (SLM), enabling gen AI capabilities to run on-premises without requiring integration with external language model providers. The SLM patch provides a bundled, local small language model (Ministral-3-3B-Instruct-Q4_K_M) and inference engine (Ollama) to run inside the Guardium environment. After installation, Guardium routes gen AI requests to the local model, which allows existing and future gen AI capabilities to operate without external AI connectivity.
For more information, see Small Language Model patch for offline generative AI.
[{"Type":"MASTER","Line of Business":{"code":"LOB76","label":"Data Platform"},"Business Unit":{"code":"BU048","label":"IBM Software"},"Product":{"code":"SSDKGA","label":"IBM Guardium Data Protection"},"ARM Category":[{"code":"a8m3p000000PCTuAAO","label":"Platform\/Installation\/Deployment"}],"ARM Case Number":"","Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"12.2.4"}]
Was this topic helpful?
Document Information
Modified date:
16 October 2026
UID
ibm17289201