NVIDIA GeForce RTX 4080 Super

NVIDIA · 16GB GDDR6X · Can run 36 models

Buy Amazon
Manufacturer NVIDIA
VRAM 16 GB
Memory Type GDDR6X
Architecture Ada Lovelace
CUDA Cores 10,240
Bandwidth 736 GB/s
TDP 320W
MSRP $999
Released Jan 31, 2024

AI Notes

The RTX 4080 Super is a high-end option with 16GB VRAM and 736 GB/s bandwidth. It comfortably runs 13B models and can handle some 30B models with aggressive quantization. The high bandwidth ensures fast token generation, making it excellent for interactive AI use.

Compatible Models

Model Parameters Best Quant VRAM Used Fit Est. Speed
Gemma 3 1B 1B Q8_0 2 GB Runs ~368 tok/s
Llama 3.2 1B 1B Q8_0 3 GB Runs ~245 tok/s
DeepSeek R1 1.5B 1.5B Q8_0 3 GB Runs ~245 tok/s
Gemma 2 2B 2B Q8_0 4 GB Runs ~184 tok/s
Llama 3.2 3B 3B Q8_0 5 GB Runs ~147 tok/s
Phi-3 Mini 3.8B 3.8B Q8_0 5.8 GB Runs ~127 tok/s
Phi-4 Mini 3.8B 3.8B Q4_K_M 4.5 GB Runs ~164 tok/s
Gemma 3 4B 4B Q4_K_M 5 GB Runs ~147 tok/s
Qwen 3 4B 4B Q4_K_M 4.5 GB Runs ~164 tok/s
DeepSeek R1 7B 7B Q8_0 9 GB Runs ~82 tok/s
Mistral 7B 7B Q8_0 9 GB Runs ~82 tok/s
Qwen 2.5 7B 7B Q8_0 9 GB Runs ~82 tok/s
Qwen 2.5 Coder 7B 7B Q8_0 9 GB Runs ~82 tok/s
DeepSeek R1 8B 8B Q4_K_M 7.5 GB Runs ~98 tok/s
Llama 3.1 8B 8B Q8_0 10 GB Runs ~74 tok/s
Qwen 3 8B 8B Q4_K_M 7.5 GB Runs ~98 tok/s
Gemma 2 9B 9B Q8_0 11 GB Runs ~67 tok/s
Gemma 3 12B 12B Q4_K_M 10.5 GB Runs ~70 tok/s
Mistral Nemo 12B 12B Q4_K_M 9.5 GB Runs ~77 tok/s
DeepSeek R1 14B 14B Q4_K_M 9.9 GB Runs ~74 tok/s
Phi-4 14B 14B Q4_K_M 9.9 GB Runs ~74 tok/s
Qwen 2.5 14B 14B Q4_K_M 9.9 GB Runs ~74 tok/s
Qwen 2.5 Coder 14B 14B Q4_K_M 12 GB Runs ~61 tok/s
Qwen 3 14B 14B Q4_K_M 12 GB Runs ~61 tok/s
Codestral 22B 22B Q4_K_M 14.7 GB Runs (tight) ~50 tok/s
StarCoder2 15B 15B Q8_0 17 GB CPU Offload ~43 tok/s
Devstral 24B 24B Q4_K_M 17 GB CPU Offload ~43 tok/s
Mistral Small 3.1 24B 24B Q4_K_M 18 GB CPU Offload ~41 tok/s
Gemma 2 27B 27B Q4_K_M 17.7 GB CPU Offload ~42 tok/s
Gemma 3 27B 27B Q4_K_M 20 GB CPU Offload ~37 tok/s
Qwen 3 30B-A3B (MoE) 30B Q4_K_M 22 GB CPU Offload ~33 tok/s
DeepSeek R1 32B 32B Q4_K_M 20.7 GB CPU Offload ~36 tok/s
Qwen 2.5 32B 32B Q4_K_M 20.7 GB CPU Offload ~36 tok/s
Qwen 2.5 Coder 32B 32B Q4_K_M 23 GB CPU Offload ~32 tok/s
Qwen 3 32B 32B Q4_K_M 23 GB CPU Offload ~32 tok/s
Command R 35B 35B Q4_K_M 22.5 GB CPU Offload ~33 tok/s
7 model(s) are too large for this hardware.