Jamba
Hybrid SSM-Transformer model by AI21 Labs combining Mamba with attention layers.
About
Jamba by AI21 Labs is a hybrid foundation model that interleaves Mamba state-space layers with Transformer attention and Mixture-of-Experts layers, delivering long-context handling up to 256K tokens with efficient inference. It was the first non-Transformer-based architecture scaled to the quality of leading models of its size, released as instruction-following open weights. Released under the Apache 2.0 license.
Reviews (0)
Leave a Review
No reviews yet. Be the first to review!
Details
- Category
- Large Language Models (LLMs)
- Price
- Free
- Platform
- Local/Desktop
- Difficulty
- Intermediate (3/5)
- License
- Apache-2.0
- Minimum VRAM
- 12 GB
- Added
- Apr 3, 2026
Related Tools
Open-source code LLM family by IBM for enterprise code generation.
High-performance open-weight MoE LLM with 671B total parameters.
State-space model for language with linear-time sequence modeling.
Family of small language models by Hugging Face for on-device use.
Open-source multimodal LLM competitive with commercial models by Shanghai AI Lab.
Open-weight LLM by Meta in 8B and 70B sizes with strong general capabilities.