architectures · ai21
Interleaved Mamba2 and full-attention transformer blocks with MoE FFN layers. Achieves Transformer quality with 50× KV-cache reduction for long documents.
| Weight file SHA-256 | 9c51c502ef60958e051825b36954a492b701dd842c4c2ee8448b125dc4bf0f17 |
|---|
| Publisher | ai21 |
|---|---|
| Origin | Israel |
| Licence | Jamba-Open-Model |
| Modality | Text |
| Pulls recorded here | 28,000 |
Registry record as of 2026-06-06. Identity, licence and parameter count are properties of a release and do not move; tier can, and the verify link above re-checks it against the live registry rather than this page.