How much VRAM does Jamba 1.5 Mini need?

About 33.0 GB at Q4_K_M with 4K of context and batch 1, rising to roughly 33.5 GB at 32K. That figure is weights plus KV cache plus a 10% activation buffer, calculated from the model's architecture rather than measured on a card.

Will Jamba 1.5 Mini run on a 40GB GPU?

Yes — at Q4_K_M and 4K context it needs about 33.0 GB, which leaves 7.0 GB spare on a A100 40G. That is the smallest card in this index that clears it comfortably; 18 of 78 do.

Which quantization of Jamba 1.5 Mini should I use?

Q4_K_M has the lowest published quality loss (3.4%), and Q4_K_M is the level most people run. All 2 levels in the index are Q4_K_M, AWQ INT4. Note this is a mixture-of-experts model — all parameters must be resident even though only a fraction are active per token, so the memory cost follows the total, not the active count.

Jamba 1.5 Mini

52B-A12B

AI21 Labs

⚠ Superseded · Prefer Ministral 3 8B Instruct

This index would suggest Ministral 3 8B Instruct instead today — more quant levels to choose from (2 → 3). The numbers below are still accurate; they are just for a model you probably should not start with. Compare the two →

AI21's hybrid Mamba-Transformer MoE: 52B total, about 12B active, 256K context. Only 4 of its 32 layers keep a KV cache, so long context is cheap — but all 52B of weights still have to be resident, about 31 GB at Q4. That is a 32 GB card or a large Mac, not the 16 GB card this entry used to claim. Sizes use the calculator's generic rates; no speed has been measured here.

Pro GPUMac / Apple Silicon

262K

Max Context

Quant Variants

GGUF Q4_K_M

Best Quality

96.6%

Accuracy Retained

Calculate VRAM Hugging Face Compare

Quantization Variants

Per-quant VRAM, quality loss, and inference speed on RTX 4090

Measured = run on this site · Estimated = not run here: calculated, or a figure this site has not verified · Community = public reports

Format	Level	BPW	VRAM	PPL Loss	Speed	Source	Actions
GGUF	Q4_K_M	4.85	31.5 GB	3.4%	—	Estimated	Calc HF
AWQ	INT4	4	26.0 GB	4.8%	—	Estimated	Calc HF

Similar models

Compare with Llama 3.1

70B

Llama 3.1 70B Instruct

Meta Llama 3.1

Pro GPUMac / Apple Silicon

33.4 GBmin VRAM98.8%accuracy

Meta's frontier 70B model. Requires 40GB+ VRAM; dual 3090 or M2 Ultra.

70B

Llama 3.3 70B Instruct

Meta Llama 3.3

Pro GPUMac / Apple Silicon

38.2 GBmin VRAM99.0%accuracy

Latest Meta 70B with improved multilingual. Drop-in upgrade from Llama 3.1 70B.

117B MoE

GPT-OSS 120B

OpenAI GPT-OSS

Pro GPUMac / Apple Silicon

58.5 GBmin VRAM100.0%accuracy

The big GPT-OSS (117B total / 5.1B active). Native MXFP4 checkpoint is ~61GB — fits one 80GB card or a 128GB unified-memory Mac. Partial offload on 24GB consumer cards is slow but works.

106B MoE

GLM-4.5-Air

Zhipu GLM-4.5

Pro GPUMac / Apple Silicon

34.8 GBmin VRAM97.6%accuracy

Zhipu's agentic/reasoning MoE (106B total / 12B active). Q4 ~64GB — the sweet spot is a 96GB+ unified Mac or 2× 48GB cards. Strong tool-calling for its class.

Running Jamba 1.5 Mini locally

At Q4_K_M and 4K of context, Jamba 1.5 Mini needs about 33.0 GB — 29.9 GB of weights, 0.06 GB of KV cache and a 3.0 GB activation buffer. The smallest card in this index that clears that comfortably is the A100 40G at 40 GB, and 18 of the 78 cards here do. These are calculated figures, not measurements: the estimate stops counting a card as comfortable at 88% of its VRAM, which is roughly the room a desktop session needs.

What longer context costs

Going from 4K to 32K adds about 0.44 GB, taking the total to 33.5 GB. That is gentler than the parameter count suggests: only 4 of 32 layers keep a KV cache that grows with context, so the cache scales at a fraction of the usual rate. The model's native window is 256K; holding all of it at Q4_K_M would need about 37 GB.

Which build to download

This index tracks 2 formats for it — GGUF, AWQ — across 2 levels. Q4_K_M carries the lowest published perplexity loss at 3.4%. No throughput has been measured for this model on this site's hardware. GGUF runs on llama.cpp and Ollama across NVIDIA, AMD and Apple silicon; AWQ and GPTQ target vLLM on CUDA and ROCm; EXL2 is ExLlamaV2 — CUDA only, and now archived, so local use rather than serving.

Common questions

How much VRAM does Jamba 1.5 Mini need?: About 33.0 GB at Q4_K_M with 4K of context and batch 1, rising to roughly 33.5 GB at 32K. That figure is weights plus KV cache plus a 10% activation buffer, calculated from the model's architecture rather than measured on a card.
Will Jamba 1.5 Mini run on a 40GB GPU?: Yes — at Q4_K_M and 4K context it needs about 33.0 GB, which leaves 7.0 GB spare on a A100 40G. That is the smallest card in this index that clears it comfortably; 18 of 78 do.
Which quantization of Jamba 1.5 Mini should I use?: Q4_K_M has the lowest published quality loss (3.4%), and Q4_K_M is the level most people run. All 2 levels in the index are Q4_K_M, AWQ INT4. Note this is a mixture-of-experts model — all parameters must be resident even though only a fraction are active per token, so the memory cost follows the total, not the active count.

Where this model fits

Sized at Q4_K_M with a 4K context window, smallest card first. Comfortable means the estimate uses at most 88% of the memory.

Runs comfortably on: A100 40G40GB · +7.0 A40 48G48GB · +15.0 L40S 48G48GB · +15.0 Radeon PRO W7900 48G48GB · +15.0 DGX Spark 64G64GB · +31.0
Tight but possible: RTX 509032GB Arc Pro B70 32G32GB Radeon AI PRO R9700 32G32GB
Just misses: Mac M3 Pro 36G+6.0GB over Mac M4 Max 36G+6.0GB over
Ships in: GGUF AWQ
Compared here: GGUF vs AWQ
Try it yourself: Size it in the calculator Generate the run command

How to actually run this

Deployment guides for this model and this class of hardware.

This page's figures change when the model or the runtime does.Last updated 2026-10-20 RSS → /feed.xml