Granite Guardian Models by IBM Research
About
The Granite Guardian family of language models is engineered to identify and mitigate risks within user prompts and responses by serving as safety guardrails. These models, developed by IBM, are adept at pinpointing issues such as social bias, hate speech, toxicity, and other harmful content, as categorized in IBM's AI Risk Atlas. Additionally, they offer specific checks for concerns related to Retrieval Augmented Generation (RAG), such as groundedness and context relevance. Available in 8B and 2B parameter sizes, Granite Guardian models can be integrated with any open or proprietary language models to bolster AI safety and responsible use practices. Notably, in testing, the Granite Guardian 3.0 8B model surpassed Meta's LlamaGuard 3 8B by improving the average F1-score by 4 points on standard risk detection benchmarks 345.
Current Variants
Use-when guidance is derived from seed capabilities, context, release, and replacement fields.
Use when the workload needs safety, 8k context, and 8B parameters.
Use when the workload needs safety, 128k context, and 8B parameters.
Use when the workload needs safety, 4k context, and 5B parameters.
Use when the workload needs safety, 128k context, and 8B parameters.
Use when the workload needs safety, 128k context, and 2B parameters.
Use when the workload needs safety, 4k context, and 2B parameters.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Granite Guardian 4.1 8B | Use when the workload needs safety, 8k context, and 8B parameters. | 2026-04 | safety8k context8B parameters | Current |
| Granite Guardian 3.3 8B | Use when the workload needs safety, 128k context, and 8B parameters. | 2025-08 | safety128k context8B parameters | Current |
| Granite Guardian 3.2 5B | Use when the workload needs safety, 4k context, and 5B parameters. | 2025-02 | safety4k context5B parameters | Current |
| Granite Guardian 3.1 8B | Use when the workload needs safety, 128k context, and 8B parameters. | 2024-12 | safety128k context8B parameters | Current |
| Granite Guardian 3.1 2B | Use when the workload needs safety, 128k context, and 2B parameters. | 2024-12 | safety128k context2B parameters | Current |
| Granite Guardian 3.0 2B | Use when the workload needs safety, 4k context, and 2B parameters. | 2024-10 | safety4k context2B parameters | Current |
Release Timeline
5 release groupsSpecifications(6 models)
| Model | Released | Context | Parameters |
|---|---|---|---|
| Granite Guardian 4.1 8B | 2026-04 | 8k | 8B |
| Granite Guardian 3.3 8B | 2025-08 | 128k | 8B |
| Granite Guardian 3.2 5B | 2025-02 | 4k | 5B |
| Granite Guardian 3.1 8B | 2024-12 | 128k | 8B |
| Granite Guardian 3.1 2B | 2024-12 | 128k | 2B |
| Granite Guardian 3.0 2B | 2024-10 | 4k | 2B |
Frequently Asked Questions
- What is Granite Guardian used for?
- Granite Guardian is used for safety and agent workflows. The family description and listed model capabilities point to those workloads as the best fit.
- How does Granite Guardian compare to Granite 4?
- Granite Guardian by IBM Research is strongest where you need safety, while Granite 4 by IBM Research is the closest related family to check for audio. Granite Guardian has 6 listed variants and reaches up to 128k context, while Granite 4 reaches up to 131k context, so compare the specs and pricing tables before choosing a production model.
- Which Granite Guardian model should I use?
- If price is the main constraint, use the pricing table first because Granite Guardian does not have complete provider pricing in the local data. For the most capable/latest local choice, evaluate Granite Guardian 3.3 8B with 128k context.
