The IBM Granite 2B and 8B models are designed to support tool-based use cases and support for retrieval augmented generation (RAG), streamlining code generation, translation and bug fixing.

tools 2b 8b

15.4K 2 weeks ago

Readme

An illustration of Ollama holding a beautiful flower with the IBM Rebus logo of the Eye, Bee and M, made by Paul Rand.

Granite dense models

The IBM Granite 2B and 8B models are text-only dense LLMs trained on over 12 trillion tokens of data, demonstrated significant improvements over their predecessors in performance and speed in IBM’s initial testing. Granite-8B-Instruct now rivals Llama 3.1 8B-Instruct across both OpenLLM Leaderboard v1 and OpenLLM Leaderboard v2 benchmarks.

They are designed to support tool-based use cases and for retrieval augmented generation (RAG), streamlining code generation, translation and bug fixing.

Parameter Sizes

2B:

ollama run granite3-dense:2b

8B:

ollama run granite3-dense:8b

Supported Languages

English, German, Spanish, French, Japanese, Portuguese, Arabic, Czech, Italian, Korean, Dutch, Chinese (Simplified)

Capabilities

  • Summarization
  • Text classification
  • Text extraction
  • Question-answering
  • Retrieval Augmented Generation (RAG)
  • Code related
  • Function-calling
  • Multilingual dialog use cases

Granite mixture of experts models

The Granite mixture of experts models are available in 1B and 3B parameter sizes designed for low latency usage.

See model page

Learn more