Qwen3-VL 30B-A3B Instruct

30B-A3B

Alibaba Qwen3-VL

Multimodal MoE with only ~3B active parameters, so it stays responsive on Apple unified memory and survives CPU offload far better than a dense 30B. Q4 ~19GB — a 24GB card holds it outright.

⬇ 406.2K HF downloads♥ 591 likesQwen/Qwen3-VL-30B-A3B-Instruct· stats from 8/14/2026

Consumer GPUPro GPUMac / Apple Silicon

41K

Max Context

Quant Variants

GGUF Q8_0

Best Quality

99.7%

Accuracy Retained

Calculate VRAM Hugging Face Compare

Quantization Variants

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

Measured = site benchmarks · Estimated = formula · Community = public reports

Format	Level	BPW	VRAM	PPL Loss	Speed	Source	Actions
GGUF	Q4_K_M	4.85	19.0 GB	2.6%	95 tok/s	Community	Calc HF
GGUF	Q3_K_M	3.87	15.2 GB	5.4%	104 tok/s	Estimated	Calc HF
GGUF	Q8_0	8.5	32.6 GB	0.3%	78 tok/s	Estimated	Calc HF
AWQ	INT4	4	17.2 GB	3.7%	112 tok/s	Estimated	Calc HF

Similar models

Compare with Qwen3-VL 8B

Qwen3-VL 8B Instruct

Alibaba Qwen3-VL

Consumer GPUMac / Apple Silicon

5.3 GBmin VRAM·99.7%accuracy

Current-generation vision-language model that still fits a single 8–12GB card at Q4 (~5.9GB). The realistic multimodal option for people without a 24GB GPU — note the vision encoder adds VRAM the KV-cache math below does not model.

27B

Gemma 3 27B IT

Google Gemma 3

Consumer GPUPro GPU

13.0 GBmin VRAM·97.2%accuracy

Gemma 3 large instruct with long context and multimodal support. Q4 ~16GB — dual-GPU or 24GB card with short ctx.

24B

Magistral Small 1.2 24B

Mistral Magistral

Consumer GPUMac / Apple Silicon

13.0 GBmin VRAM·99.4%accuracy

Mistral's reasoning model on a Mistral Small 3.2 base, with reasoning wrapped in [THINK] tags. Q4 ~14GB puts explicit chain-of-thought on a 16GB card — the level below the 70B-class reasoners most guides assume.

36B

Seed-OSS 36B Instruct

ByteDance Seed

Consumer GPUPro GPU

17.4 GBmin VRAM·98.7%accuracy

Dense 36B with a native 512K context. Q4 weights are ~22GB, which lands on a 24GB card or a 2×16GB split — but the context is the real cost: 512K of KV cache is roughly 128GB on its own, so budget context first and weights second.

How to actually run this

Deployment guides for this model and this class of hardware.