ibm-granite/granite-4.0-h-small

Text Generation • 32B • Updated Nov 3, 2025 • 31k • 283

Note Granite 4.0 is open-sourced under the Apache 2.0 license and is the first model family to receive ISO 42001 certification for AI safety and transparency. Please take a look at the other 4.0 models available: https://huggingface.co/collections/ibm-granite/granite-40-nano-language-models https://huggingface.co/collections/ibm-granite/granite-40-language-models Reasoning is not yet available but expected in the near future.

ibm-granite/granite-3.3-8b-instruct

Text Generation • 8B • Updated May 12, 2025 • 63k • 144

Note REASONING MODEL (namely, for coding tasks) IMPORTANT NOTE: It is possible that some of the prompt data (user input) on which this model was trained was synthesized with proprietary models, however, as inferred from the transparency report, all model output was either (1) taken from permissive datasets (many of which were filtered meta-datasets) or (2) synthesized with internal permissive models. Thank you to IBM for making one of the first quality instruct models with high ethics and legality.

ibm-granite/granite-3.0-8b-instruct

Text Generation • 8B • Updated Dec 19, 2024 • 16.8k • 204

Note INSTRUCT MODEL (namely, for coding tasks) IMPORTANT NOTE: It is possible that some of the prompt data (user input) on which this model was trained was synthesized with proprietary models, however, as inferred from the transparency report, all model output was either (1) taken from permissive datasets (many of which were filtered meta-datasets) or (2) synthesized with internal permissive models. Thank you to IBM for making one of the first quality instruct models with high ethics and legality.

alea-institute/kl3m-003-1.7b

Text Generation • 2B • Updated Apr 10, 2025 • 33 • 3

Note BASE MODEL There are many other checkpoints. https://huggingface.co/alea-institute/models

alea-institute/kl3m-003-3.7b

Text Generation • 4B • Updated Apr 10, 2025 • 17 • 2

Note BASE MODEL There are many other checkpoints. https://huggingface.co/alea-institute/models

common-pile/comma-v0.1-1t

7B • Updated Jun 6, 2025 • 15 • 25

Note BASE MODEL (1 trillion tokens)

common-pile/comma-v0.1-2t

7B • Updated Jun 6, 2025 • 9.21k • 32

Note BASE MODEL (2 trillion tokens)

PleIAs/Pleias-350m-Preview

0.4B • Updated Feb 14, 2025 • 165 • 24

Note BASE MODEL (preview)

PleIAs/Pleias-3b-Preview

3B • Updated Dec 5, 2024 • 70 • 32

Note BASE MODEL (preview)

PleIAs/Pleias-Pico

0.4B • Updated Feb 14, 2025 • 38 • 34

Note INSTRUCT/RAG MODEL

PleIAs/Pleias-Nano

1B • Updated Dec 5, 2024 • 24 • 37

Note INSTRUCT/RAG MODEL

ibm-granite/granite-3.0-8b-base

Text Generation • 8B • Updated Dec 19, 2024 • 1.68k • 24

Note BASE MODEL Supports FIM. Awesome!

ibm-granite/granite-3.3-8b-base

Text Generation • 8B • Updated May 8, 2025 • 5.52k • 24

Note BASE MODEL Supports FIM. Awesome!