Skip to main content

IBM · Text

IBM Granite models: Versions, specs, GPU requirements & hosting

IBM Granite provides enterprise-focused language checkpoints across compact and larger deployment sizes with permissive release-level licensing.

Current releases Browse engineering services Compare models
Technical deployment reference illustration for IBM Granite
Reference deployment flow; exact architecture facts are listed separately.

Overview and history

IBM Granite is an official IBM model family represented here through independently verified release, checkpoint, license, runtime and hardware evidence. IBM Granite provides enterprise-focused language checkpoints across compact and larger deployment sizes with permissive release-level licensing. Current public releases are grouped by the General tracks so sibling sizes and specialist branches are not presented as false chronological generations.

FusionPointAI currently tracks 12 certified public IBM Granite releases. The catalogue preserves release dates and evaluates Current, Recent, Previous and Legacy status within each model track rather than by database insertion order.

Enterprise deployment focus
Multiple model sizes
Broad runtime support

Architecture and model range

The verified public range spans 3B to 30B total parameters and context windows up to 131,072 tokens. Each release states whether it is dense or mixture-of-experts and records active parameters separately when the official source publishes them.

Public releases12
Model tracks1
Current generationGranite 4.2 30B
Last verified2026-09-03T00:00:00.000000Z

Generations, tracks and variants

General

Granite 3.3

Previous generation
Granite 3.3 8B

8.0B parameters · 131,072 context

Dense Dense Transformer

granite-4-2-3b

Recent generation 5 sibling variants
granite-4.2-3b

Not published by manufacturer parameters · N/A / Not directly comparable context

Base
granite-4.2-3b-fp8

Not published by manufacturer parameters · N/A / Not directly comparable context

Quantization
granite-4.2-3b-nvfp4

Not published by manufacturer parameters · N/A / Not directly comparable context

Quantization
granite-4.2-3b-mxfp4

Not published by manufacturer parameters · N/A / Not directly comparable context

Quantization
granite-4.2-3b-GGUF

Not published by manufacturer parameters · N/A / Not directly comparable context

Quantization

Granite 4.1

Recent generation
Granite 4.1 30B

28.9B parameters · 131,072 context

Dense GraniteForCausalLM

granite-4-1-30b

Recent generation 2 sibling variants
granite-4.1-30b-GGUF

Not published by manufacturer parameters · N/A / Not directly comparable context

Quantization
granite-4.1-30b-fp8

Not published by manufacturer parameters · N/A / Not directly comparable context

Quantization

Granite 4.2

Current generation 3 sibling variants
Granite 4.2 8B

8.0B parameters · 131,072 context

Reasoning · Dense GraniteForCausalLM
Granite 4.2 3B

3.0B parameters · 131,072 context

Reasoning · Dense GraniteForCausalLM
Granite 4.2 30B

30.0B parameters · 131,072 context

Reasoning · Dense GraniteForCausalLM

Deployment ecosystem

Choose a specific IBM Granite release and certified checkpoint before sizing hardware. FusionPointAI rejects unknown runtime configurations, calculates resident memory from total model weights, then distinguishes globally compatible GPUs from currently rentable inventory.

Validate the exact checkpoint license, runtime version, precision kernel, context length, concurrency and multi-GPU topology. Active MoE parameters affect compute, but resident model memory continues to use total stored weights.

Common use cases

Business assistants
Document processing
Private inference

License overview

Licenses are certified per release, not inferred from the provider. Review the release page before commercial deployment because terms can differ between generations and variants.

Commercial-use status is a source-backed catalogue classification, not legal advice. Follow the linked official license and attribution requirements.

Release timeline

ibm/granite/granite-41-30b-gguf @ a4a42312a56a3dff29f67d89361b375e5f1efa375fc7607cc1417ea127727e3c

Official source

ibm/granite/granite-41-30b-fp8 @ 2006d06729657c5b131fc9971f6d5dabf272bdbd3293c646dd4e134414edaeae

Official source

ibm/granite/granite-4-1-30b @ c9d5fa2930a611bb481573dc233b242ee8f2b45f60d32c68ce6bb1bd80e1be0c

Official source

ibm/granite/granite-42-3b-gguf @ 83f847c71f7e86c289128785baf06d0ec25033683ea92012d230707afbb6f01e

Official source

ibm/granite/granite-42-3b-nvfp4 @ e9aa554262fef214b1a003c3a759526ab710f48e3517b08308a9e64bc67e557a

Official source

ibm/granite/granite-42-3b-mxfp4 @ d523d0dbb104704c5418a46a1b77d8586f5dd7e866b607c11584366178f689fc

Official source

ibm/granite/granite-42-3b-fp8 @ 6f3e2b6caad002f9daf7726d085987d8a97d1993216336c9dddc7fbe8f642b7f

Official source

ibm/granite/granite-3-3-8b @ 180c386c903c1543d8c3c4f94e27491c4b6cb66fbece18abcbd198d01e320468

Official source

ibm/granite/granite-4-2-30b @ 37162f41301604396463c91856e5c7bdfaf69aaab7a36920fee97fedd14fa694

Official source

ibm/granite/granite-4-2-8b @ cab33b9e1d79489645311c751bc727a0c725f637793648e9c33bcccf04de3b78

Official source

ibm/granite/granite-4-2-3b @ dd9a1fe61833ecf5e9963237b60da9469bf3d5cf750bd93cde2c2997bb74da98

Official source

ibm/granite/granite-42-3b @ e01442b49f274eabb1750952a68ef40b032c48a6bff7dc8d1120c4ba80881aa1

Official source

ibm-granite/granite-4.2-3b @ 2026-08-25T18:42:47Z

Official source

Frequently asked questions

Which IBM Granite release is current?

Current releases are Granite 4.2 30B, Granite 4.2 8B, Granite 4.2 3B, evaluated within their respective tracks.

Can IBM Granite be self-hosted?

Only releases with a verified official checkpoint and supported runtime configuration are presented as deployable.

Which GPU should run IBM Granite?

GPU selection depends on the exact variant, checkpoint precision, context length and concurrency; use the release hardware recommendations rather than the family name alone.

Official references

Last verified: 2026-09-03