Skip to content

NVIDIA RTX A6000

NVIDIA · 48GB GDDR6 · Can run 92 models

Buy Amazon
Manufacturer NVIDIA
VRAM 48 GB
Memory Type GDDR6
Architecture Ampere
CUDA Cores 10,752
Tensor Cores 336
Bandwidth 768 GB/s
TDP 300W
MSRP $4,650
Released Dec 3, 2020

AI Notes

The RTX A6000 is a workstation GPU with a massive 48GB of GDDR6 VRAM — the most of any single Ampere GPU. It can load 30B models at full precision and run 70B models with quantization. While expensive new, used prices have dropped significantly, making it one of the best options for running large models on a single GPU.

Compatible Models

Model Parameters Best Quant VRAM Used Fit Est. Speed
Qwen 3 0.6B 600M Q4_K_M 2.5 GB Runs ~307 tok/s
Qwen 3.5 0.8B 800M Q4_K_M 1.5 GB Runs ~512 tok/s
Gemma 3 1B 1B Q8_0 2 GB Runs ~384 tok/s
Llama 3.2 1B 1B Q8_0 3 GB Runs ~256 tok/s
DeepSeek R1 1.5B 1.5B Q8_0 3 GB Runs ~256 tok/s
SmolLM2 1.7B 1.7B Q8_0 2.7 GB Runs ~284 tok/s
Gemma 2 2B 2B Q8_0 4 GB Runs ~192 tok/s
Gemma 3n E2B 2B Q4_K_M 3.3 GB Runs ~233 tok/s
Gemma 4 E2B 2B Q4_K_M 4 GB Runs ~192 tok/s
Qwen 3.5 2B 2B Q4_K_M 3 GB Runs ~256 tok/s
Llama 3.2 3B 3B Q8_0 5 GB Runs ~154 tok/s
StarCoder2 3B 3B Q4_K_M 3.5 GB Runs ~219 tok/s
Phi-3 Mini 3.8B 3.8B Q8_0 5.8 GB Runs ~132 tok/s
Phi-4 Mini 3.8B 3.8B Q4_K_M 4.5 GB Runs ~171 tok/s
Gemma 3 4B 4B Q4_K_M 5 GB Runs ~154 tok/s
Gemma 3n E4B 4B Q4_K_M 4.5 GB Runs ~171 tok/s
Gemma 4 E4B 4B Q4_K_M 6 GB Runs ~128 tok/s
Qwen 3 4B 4B Q4_K_M 4.5 GB Runs ~171 tok/s
Qwen 3.5 4B 4B Q4_K_M 4.5 GB Runs ~171 tok/s
Yi 1.5 6B 6B Q4_K_M 5 GB Runs ~154 tok/s
Codestral Mamba 7B 7B Q4_K_M 6.9 GB Runs ~111 tok/s
DeepSeek R1 7B 7B Q8_0 9 GB Runs ~85 tok/s
Falcon 3 7B 7B Q4_K_M 6.8 GB Runs ~113 tok/s
InternLM 2.5 7B 7B Q4_K_M 5.5 GB Runs ~140 tok/s
Mistral 7B 7B Q8_0 9 GB Runs ~85 tok/s
OpenChat 3.5 7B 7B Q4_K_M 6.9 GB Runs ~111 tok/s
Qwen 2.5 7B 7B Q8_0 9 GB Runs ~85 tok/s
Qwen 2.5 Coder 7B 7B Q8_0 9 GB Runs ~85 tok/s
Qwen 2.5 VL 7B 7B Q4_K_M 7 GB Runs ~110 tok/s
StarCoder2 7B 7B Q4_K_M 5.5 GB Runs ~140 tok/s
WizardLM 2 7B 7B Q4_K_M 6.9 GB Runs ~111 tok/s
Aya Expanse 8B 8B Q4_K_M 6.5 GB Runs ~118 tok/s
Cogito 8B 8B Q4_K_M 7.5 GB Runs ~102 tok/s
DeepSeek R1 8B 8B Q4_K_M 7.5 GB Runs ~102 tok/s
Dolphin 3 8B 8B Q4_K_M 6 GB Runs ~128 tok/s
Granite 3.3 8B 8B Q8_0 10 GB Runs ~77 tok/s
Llama 3.1 8B 8B Q8_0 10 GB Runs ~77 tok/s
Nemotron 3 Nano 8B 8B Q4_K_M 7.5 GB Runs ~102 tok/s
Nous Hermes 2 8B 8B Q4_K_M 6 GB Runs ~128 tok/s
Qwen 3 8B 8B Q4_K_M 7.5 GB Runs ~102 tok/s
Gemma 2 9B 9B Q8_0 11 GB Runs ~70 tok/s
Qwen 3.5 9B 9B Q4_K_M 7.5 GB Runs ~102 tok/s
Yi 1.5 9B 9B Q4_K_M 6.5 GB Runs ~118 tok/s
Yi Coder 9B 9B Q4_K_M 8 GB Runs ~96 tok/s
Falcon 3 10B 10B Q4_K_M 8.5 GB Runs ~90 tok/s
Llama 3.2 Vision 11B 11B Q4_K_M 8.5 GB Runs ~90 tok/s
Gemma 3 12B 12B Q4_K_M 10.5 GB Runs ~73 tok/s
Mistral Nemo 12B 12B Q4_K_M 9.5 GB Runs ~81 tok/s
DeepSeek R1 14B 14B Q4_K_M 9.9 GB Runs ~78 tok/s
Phi-4 14B 14B Q4_K_M 9.9 GB Runs ~78 tok/s
Phi-4 Reasoning 14B 14B Q4_K_M 11 GB Runs ~70 tok/s
Qwen 2.5 14B 14B Q4_K_M 9.9 GB Runs ~78 tok/s
Qwen 2.5 Coder 14B 14B Q4_K_M 12 GB Runs ~64 tok/s
Qwen 3 14B 14B Q4_K_M 12 GB Runs ~64 tok/s
StarCoder2 15B 15B Q8_0 17 GB Runs ~45 tok/s
InternLM 2.5 20B 20B Q4_K_M 12 GB Runs ~64 tok/s
Codestral 22B 22B Q4_K_M 14.7 GB Runs ~52 tok/s
Devstral 24B 24B Q4_K_M 17 GB Runs ~45 tok/s
Magistral Small 24B 24B Q4_K_M 17 GB Runs ~45 tok/s
Mistral Small 3.1 24B 24B Q4_K_M 18 GB Runs ~43 tok/s
Gemma 4 26B 26B Q4_K_M 20 GB Runs ~38 tok/s
Gemma 2 27B 27B Q4_K_M 17.7 GB Runs ~43 tok/s
Gemma 3 27B 27B Q4_K_M 20 GB Runs ~38 tok/s
Qwen 3.5 27B 27B Q4_K_M 19 GB Runs ~40 tok/s
Qwen 3 30B-A3B (MoE) 30B Q4_K_M 22 GB Runs ~35 tok/s
Gemma 4 31B 31B Q4_K_M 22 GB Runs ~35 tok/s
Aya Expanse 32B 32B Q4_K_M 22 GB Runs ~35 tok/s
Cogito 32B 32B Q4_K_M 21.5 GB Runs ~36 tok/s
DeepSeek R1 32B 32B Q4_K_M 20.7 GB Runs ~37 tok/s
Qwen 2.5 32B 32B Q4_K_M 20.7 GB Runs ~37 tok/s
Qwen 2.5 Coder 32B 32B Q4_K_M 23 GB Runs ~33 tok/s
Qwen 3 32B 32B Q4_K_M 23 GB Runs ~33 tok/s
QwQ 32B 32B Q4_K_M 21.5 GB Runs ~36 tok/s
WizardCoder 33B 33B Q4_K_M 22 GB Runs ~35 tok/s
Nous Hermes 2 34B 34B Q4_K_M 19 GB Runs ~40 tok/s
Yi 1.5 34B 34B Q4_K_M 21 GB Runs ~37 tok/s
Command R 35B 35B Q4_K_M 22.5 GB Runs ~34 tok/s
Qwen 3.5 35B A3B 35B Q4_K_M 12 GB Runs ~64 tok/s
Dolphin Mixtral 8x7B 47B Q4_K_M 26 GB Runs ~30 tok/s
Mixtral 8x7B 47B Q4_K_M 29.7 GB Runs ~26 tok/s
Cogito 70B 70B Q4_K_M 43 GB Runs (tight) ~18 tok/s
DeepSeek R1 70B 70B Q4_K_M 43.5 GB Runs (tight) ~18 tok/s
Llama 3.1 70B 70B Q4_K_M 43.5 GB Runs (tight) ~18 tok/s
Llama 3.3 70B 70B Q4_K_M 43.5 GB Runs (tight) ~18 tok/s
Qwen 2.5 72B 72B Q4_K_M 44.7 GB Runs (tight) ~17 tok/s
Qwen 2.5 VL 72B 72B Q4_K_M 41 GB Runs (tight) ~19 tok/s
Llama 3.2 Vision 90B 90B Q4_K_M 50 GB CPU Offload ~5 tok/s
Command R+ 104B 104B Q4_K_M 57 GB CPU Offload ~4 tok/s
Llama 4 Scout (109B/17B active) 109B Q4_K_M 72 GB CPU Offload ~3 tok/s
Command A 111B 111B Q4_K_M 61 GB CPU Offload ~4 tok/s
Devstral 2 123B 123B Q4_K_M 67 GB CPU Offload ~3 tok/s
Mistral Large 2 123B 123B Q4_K_M 67 GB CPU Offload ~3 tok/s
13 model(s) are too large for this hardware.