Question 1

How much VRAM does AI21-Jamba-Large-1.5 need?

Accepted Answer

AI21-Jamba-Large-1.5 (398.6B) requires approximately 890.8401705980301 GB of VRAM at full precision (FP16), 445.4 GB at 8-bit quantization (Q8), and 222.7 GB at 4-bit quantization (Q4). The actual VRAM usage may be slightly higher due to KV cache and runtime overhead.

Question 2

What GPUs can run AI21-Jamba-Large-1.5?

Accepted Answer

We currently have no compatible GPUs indexed for AI21-Jamba-Large-1.5. This may be because the model's VRAM requirements exceed available GPUs in our database, or VRAM data hasn't been computed yet.

Question 3

What is the cheapest GPU to run AI21-Jamba-Large-1.5?

Accepted Answer

We don't currently have cloud pricing data for GPUs compatible with AI21-Jamba-Large-1.5. Cloud pricing data is updated regularly — check back soon.

Question 4

Can AI21-Jamba-Large-1.5 run on consumer hardware?

Accepted Answer

AI21-Jamba-Large-1.5 requires approximately 890.8401705980301 GB of VRAM at FP16, and even with 4-bit quantization (Q4) it needs about 222.7 GB. This exceeds the 24 GB available on the highest-end consumer GPU (RTX 4090), so AI21-Jamba-Large-1.5 requires datacenter or cloud GPUs.

Question 5

What quantization options are available for AI21-Jamba-Large-1.5?

Accepted Answer

AI21-Jamba-Large-1.5 supports the following quantization levels: FP16 (full precision): ~890.8401705980301 GB, Q8 (8-bit): ~445.4 GB, Q6 (6-bit): ~334.1 GB, Q4 (4-bit): ~222.7 GB. Lower quantization reduces VRAM usage at the cost of some quality loss. Q8 offers a good balance between quality and memory savings, while Q4 provides maximum compression.

AI21-Jamba-Large-1.5

VRAM Requirements

Compatible GPUs

AI21-Jamba-Large-1.5 FAQ

VRAM requirements for AI21-Jamba-Large-1.5 at different quantization levels
Quantization	VRAM Required
FP16	890.8 GB
Q8	445.4 GB
Q6	334.1 GB
Q4	222.7 GB