Overview
Granite 3.3 continues IBM’s Granite family of open models aimed at enterprise use, with attention to data provenance and licensing.
Highlights
- 2B and 8B sizes
- 128K-token context window
- Optional reasoning mode
- Apache 2.0 license
Typical uses
Enterprise RAG, summarization, classification and tool-calling agents.
Things to consider
- Smaller community than Llama or Qwen.
- Download sizes refer to the default 4-bit quantized versions in the Ollama library; full-precision weights are larger.
Specifications are taken from the model card on Hugging Face, the Ollama library and the provider’s announcement. Benchmark results depend on evaluation settings; see how AI benchmarks work.