Wizard Vicuna 30B Uncensored GPTQ — Hardware Requirements & GPU Compatibility
Chat
Wizard Vicuna 30B Uncensored GPTQ is a 32.5B-parameter open language model from TheBloke in the Vicuna family. It supports a context window of up to 2,048 tokens. At FP16 it needs about 71.57 GB of VRAM — see which GPUs and Macs can run it below.
181 downloads
608 likesSpecifications
- Publisher
- TheBloke
- Family
- Vicuna
- Parameters
- 32.5B
- Architecture
- LlamaForCausalLM
- Context Length
- 2,048 tokens
- Vocabulary Size
- 32,000
- Release Date
- 2023-09-27
- License
- Other
Get Started
HuggingFace
How Much VRAM Does Wizard Vicuna 30B Uncensored GPTQ Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| FP16 | 16.00 | 71.6 GB | — | 65.07 GB | Full half-precision — baseline for inference |
Which GPUs Can Run Wizard Vicuna 30B Uncensored GPTQ?
FP16 · 71.6 GBWizard Vicuna 30B Uncensored GPTQ (FP16) requires 71.6 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 94+ GB is recommended. No single GPU has enough memory — multi-GPU or cluster setups are needed.
Which Devices Can Run Wizard Vicuna 30B Uncensored GPTQ?
FP16 · 71.6 GB5 devices with unified memory can run Wizard Vicuna 30B Uncensored GPTQ, including NVIDIA DGX H100, NVIDIA DGX A100 640GB.
Related Models
Frequently Asked Questions
- How much VRAM does Wizard Vicuna 30B Uncensored GPTQ need?
Wizard Vicuna 30B Uncensored GPTQ requires 71.6 GB of VRAM at FP16.
VRAM = Weights + KV Cache + Overhead
Weights = 32.5B × 16 bits ÷ 8 = 65.1 GB
KV Cache + Overhead ≈ 6.5 GB (at 2K context + ~0.3 GB framework)
VRAM usage by quantization
FP1671.6 GB- Can NVIDIA GeForce RTX 5090 run Wizard Vicuna 30B Uncensored GPTQ?
No — Wizard Vicuna 30B Uncensored GPTQ requires at least 71.6 GB at FP16, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
- Can I run Wizard Vicuna 30B Uncensored GPTQ on a Mac?
Wizard Vicuna 30B Uncensored GPTQ requires at least 71.6 GB at FP16, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run Wizard Vicuna 30B Uncensored GPTQ locally?
Yes — Wizard Vicuna 30B Uncensored GPTQ can run locally on consumer hardware. At FP16 quantization it needs 71.6 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Wizard Vicuna 30B Uncensored GPTQ?
At FP16, Wizard Vicuna 30B Uncensored GPTQ can reach ~41 tok/s on AMD Instinct MI300X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: AMD Instinct MI300X → 5300 ÷ 71.6 × 0.55 = ~41 tok/s
Estimated speed at FP16 (71.6 GB)
~41 tok/s~30 tok/s~25 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Wizard Vicuna 30B Uncensored GPTQ?
At FP16, the download is about 65.07 GB.
- Which GPUs can run Wizard Vicuna 30B Uncensored GPTQ?
No single consumer GPU has enough VRAM to run Wizard Vicuna 30B Uncensored GPTQ at FP16 (71.6 GB). Multi-GPU or professional hardware is required.
- Which devices can run Wizard Vicuna 30B Uncensored GPTQ?
5 devices with unified memory can run Wizard Vicuna 30B Uncensored GPTQ at FP16 (71.6 GB), including Mac Pro M2 Ultra (192 GB), Mac Studio M2 Ultra (192 GB), Mac Studio M4 Max (128 GB), NVIDIA DGX A100 640GB. Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.