Skip to content

AMD Radeon RX 7900 XTX

AMD · 24GB GDDR6 · Can run 50 models

Buy Amazon
Manufacturer AMD
VRAM 24 GB
Memory Type GDDR6
Architecture RDNA 3
Stream Procs 6,144
Bandwidth 960 GB/s
TDP 355W
MSRP $999
Released Dec 13, 2022

AI Notes

The RX 7900 XTX is AMD's flagship with 24GB of GDDR6 VRAM, making it competitive with the RTX 4090 for model loading capacity. It can run 13B models at full precision and 30B+ models with quantization. ROCm support for Ollama is maturing, though NVIDIA CUDA still offers broader compatibility.

Compatible Models

Model Parameters Best Quant VRAM Used Fit Est. Speed
Qwen 3 0.6B 600M Q4_K_M 2.5 GB Runs ~384 tok/s
Gemma 3 1B 1B Q8_0 2 GB Runs ~480 tok/s
Llama 3.2 1B 1B Q8_0 3 GB Runs ~320 tok/s
DeepSeek R1 1.5B 1.5B Q8_0 3 GB Runs ~320 tok/s
Gemma 2 2B 2B Q8_0 4 GB Runs ~240 tok/s
Gemma 3n E2B 2B Q4_K_M 3.3 GB Runs ~291 tok/s
Llama 3.2 3B 3B Q8_0 5 GB Runs ~192 tok/s
Phi-3 Mini 3.8B 3.8B Q8_0 5.8 GB Runs ~166 tok/s
Phi-4 Mini 3.8B 3.8B Q4_K_M 4.5 GB Runs ~213 tok/s
Gemma 3 4B 4B Q4_K_M 5 GB Runs ~192 tok/s
Gemma 3n E4B 4B Q4_K_M 4.5 GB Runs ~213 tok/s
Qwen 3 4B 4B Q4_K_M 4.5 GB Runs ~213 tok/s
DeepSeek R1 7B 7B Q8_0 9 GB Runs ~107 tok/s
Falcon 3 7B 7B Q4_K_M 6.8 GB Runs ~141 tok/s
Mistral 7B 7B Q8_0 9 GB Runs ~107 tok/s
Qwen 2.5 7B 7B Q8_0 9 GB Runs ~107 tok/s
Qwen 2.5 Coder 7B 7B Q8_0 9 GB Runs ~107 tok/s
Qwen 2.5 VL 7B 7B Q4_K_M 7 GB Runs ~137 tok/s
Cogito 8B 8B Q4_K_M 7.5 GB Runs ~128 tok/s
DeepSeek R1 8B 8B Q4_K_M 7.5 GB Runs ~128 tok/s
Llama 3.1 8B 8B Q8_0 10 GB Runs ~96 tok/s
Nemotron 3 Nano 8B 8B Q4_K_M 7.5 GB Runs ~128 tok/s
Qwen 3 8B 8B Q4_K_M 7.5 GB Runs ~128 tok/s
Gemma 2 9B 9B Q8_0 11 GB Runs ~87 tok/s
Falcon 3 10B 10B Q4_K_M 8.5 GB Runs ~113 tok/s
Llama 3.2 Vision 11B 11B Q4_K_M 8.5 GB Runs ~113 tok/s
Gemma 3 12B 12B Q4_K_M 10.5 GB Runs ~91 tok/s
Mistral Nemo 12B 12B Q4_K_M 9.5 GB Runs ~101 tok/s
DeepSeek R1 14B 14B Q4_K_M 9.9 GB Runs ~97 tok/s
Phi-4 14B 14B Q4_K_M 9.9 GB Runs ~97 tok/s
Phi-4 Reasoning 14B 14B Q4_K_M 11 GB Runs ~87 tok/s
Qwen 2.5 14B 14B Q4_K_M 9.9 GB Runs ~97 tok/s
Qwen 2.5 Coder 14B 14B Q4_K_M 12 GB Runs ~80 tok/s
Qwen 3 14B 14B Q4_K_M 12 GB Runs ~80 tok/s
StarCoder2 15B 15B Q8_0 17 GB Runs ~56 tok/s
Codestral 22B 22B Q4_K_M 14.7 GB Runs ~65 tok/s
Devstral 24B 24B Q4_K_M 17 GB Runs ~56 tok/s
Magistral Small 24B 24B Q4_K_M 17 GB Runs ~56 tok/s
Mistral Small 3.1 24B 24B Q4_K_M 18 GB Runs ~53 tok/s
Gemma 2 27B 27B Q4_K_M 17.7 GB Runs ~54 tok/s
Gemma 3 27B 27B Q4_K_M 20 GB Runs ~48 tok/s
Qwen 3 30B-A3B (MoE) 30B Q4_K_M 22 GB Runs (tight) ~44 tok/s
Cogito 32B 32B Q4_K_M 21.5 GB Runs (tight) ~45 tok/s
DeepSeek R1 32B 32B Q4_K_M 20.7 GB Runs (tight) ~46 tok/s
Qwen 2.5 32B 32B Q4_K_M 20.7 GB Runs (tight) ~46 tok/s
QwQ 32B 32B Q4_K_M 21.5 GB Runs (tight) ~45 tok/s
Command R 35B 35B Q4_K_M 22.5 GB Runs (tight) ~43 tok/s
Qwen 2.5 Coder 32B 32B Q4_K_M 23 GB CPU Offload ~13 tok/s
Qwen 3 32B 32B Q4_K_M 23 GB CPU Offload ~13 tok/s
Mixtral 8x7B 47B Q4_K_M 29.7 GB CPU Offload ~10 tok/s
19 model(s) are too large for this hardware.