Overview

Jamba mixes Transformer attention layers with Mamba state-space layers, which keeps memory use low for long inputs. Jamba Mini 1.6 targets enterprise RAG and grounded answers.

Highlights

  • 52B total, 12B active parameters
  • Hybrid Transformer-Mamba architecture
  • 256K-token context
  • Strong long-context efficiency

Typical uses

Long-document question answering and enterprise RAG.

Things to consider

  • Custom license with revenue thresholds.

Specifications are taken from the model card on Hugging Face and the provider’s announcement. Benchmark results depend on evaluation settings; see how AI benchmarks work.