/ Models Index / Jamba 1.5 Large

Jamba 1.5 Large

Architecture · AI21 · AI21 Labs · Released Aug 2024
B-

Good (B-) — Capability Grade

Jamba 1.5 Large uses a hybrid State-Space-Model + Transformer architecture (Mamba-Transformer). Strong long-context efficiency. Mid-tier reasoning. Israeli-headquartered; enterprise focus.

75
Composite / 100
/ Subscore Breakdown · 6 Capability Dimensions

Where this grade comes from.

General Reasoning
B-
Code Generation
C+
Math & STEM
C+
Tool Use & Agency
B-
Multimodal
D
Safety & Alignment
B
/ Key Events & Disclosures

Release timeline & positioning.

  • Released Aug 2024
  • Hybrid SSM-Transformer architecture
  • Israeli-headquartered
  • Strong long-context efficiency

/ Best for

Long-context workloads where SSM architecture's throughput efficiency provides operational advantage.

/ Watch out for

Mid-tier general capability. Reserve for specific long-context use cases where Jamba's architecture advantages dominate.