Granite trust, safety and security

Open the AI “black box” with IBM Granite. Build and deploy AI solutions with confidence.

Minimalist illustration of cubes in shades of blue and green

Responsible AI matters

Enterprise AI demands enterprise-grade trust. With some models trained on pirated data or producing biased outputs, it’s easy to see why it matters. IBM® Granite® models are built with security, safety and governance at their core, giving you the confidence to build responsible AI.

Learn more about responsible AI

Built in the open

True open source development
Unlike many "open" models that release only model weights, IBM Granite provides extensive transparency, including training data sources, methodologies and architectural decisions under a permissive Apache 2.0 license.
Model Openness Framework
IBM Granite qualifies as a Class III Open Model on the Linux® Foundation’s framework classifying the completeness and openness of models, an achievement not met by many model providers.
Industry-leading transparency
IBM has achieved a top five rank on Stanford’s Foundation Model Transparency Index, outperforming most major model providers.
Illustration with a pattern of colored squares in shades of blue, green, and gray

IBM further strengthens Granite for enterprise deployment with HackerOne

IBM and HackerOne started a bug bounty program for Granite with up to USD 100,000 in bounty payouts for identifying successful jailbreaks in enterprise-like settings, supporting businesses scaling AI workflows.

Learn about the program
Enterprise grade trust
Abstract illustration with geometric shapes in different colors, including green, blue, and white
Security

Granite models feature digital signatures to confirm their authenticity and specialized stress-testing programs that mimic real-world threats to uncover potential weaknesses.

Illustration of a black ball on an abstract geometric shape on a green background
Safety

Granite models go through rigorous safety testing, outperforming similar models. IBM also open sourced a benchmark to test the safety and bias detection of any AI model.

Illustration with various shades of green forming a geometric pattern with a white circle in the lower right corner
Governance

The IBM Data Management Framework Lakehouse securely manages 2.7 petabytes of training data through a robust governance process with metadata tracking and license controls.

Granite Guardian

Building on the robust safety and security features natively built into Granite models, Granite Guardian adds a layer of specialized safeguards. The guardrail models detect risks in prompts and responses from any AI model, helping to ensure safe and responsible deployment.

Download Granite Guardian Granite Guardian Docs

Comprehensive risk detection

Detects and mitigates hallucinations, harmful content, bias, jailbreaking attempts and RAG quality and accuracy issues across prompts, responses and agentic workflows.

Industry-leading performance

Granite Guardian models hold six of the top 10 spots on the GuardBench leaderboard, measuring how well guardrail models can detect harmful and hallucinated content as well as attempts to jailbreak safety controls.

Illustration with a pattern of colored blocks in shades of gray, green and blue, with a black door lock in the center