AI models and the PCs that run them
Language models
| Model | Parameters | Memory at Q4 |
|---|---|---|
| DeepSeek-R1-Distill-Qwen-7B | 8B | 7.6GB |
| DeepSeek-R1-Distill-Qwen-14B | 15B | 12.5GB |
| DeepSeek-R1-Distill-Qwen-32B | 33B | 25GB |
| DeepSeek-R1-Distill-Llama-70B | 71B | 51.5GB |
| DeepSeek-R1 | 671B | 469.8GB |
| DeepSeek-V3 | 671B | 469.8GB |
| GLM-5.2 | 753B | 527GB |
| Gemma 3 4B IT | 4B | 4.8GB |
| Gemma 3 12B IT | 12B | 10.4GB |
| Gemma 3 27B IT | 27B | 20.8GB |
| Kimi-K2-Instruct | 1000B | 699.2GB |
| Llama 3.2 3B Instruct | 3.21B | 4.2GB |
| Llama 3.1 8B Instruct | 8B | 7.6GB |
| Llama 3.1 70B Instruct | 70B | 50.8GB |
| Llama 3.3 70B Instruct | 70B | 50.8GB |
| Llama 4 Scout 17B-16E Instruct | 109B | 78GB |
| Llama 4 Maverick 17B-128E Instruct | 400B | 280.9GB |
| Mistral Small 3 (Mistral-Small-24B-Instruct-2501) | 24B | 18.7GB |
| Devstral Small 1.1 (Devstral-Small-2507) | 24B | 18.7GB |
| Mistral Medium 3.5 128B | 128B | 91.2GB |
| Phi-4 | 14B | 11.8GB |
| Qwen3-4B | 4B | 4.8GB |
| Qwen3-8B | 8.2B | 7.7GB |
| Qwen3-14B | 14.8B | 12.3GB |
| Qwen3-30B-A3B | 30.5B | 23.3GB |
| Qwen3-32B | 32.8B | 24.9GB |
| Qwen3-235B-A22B | 235B | 165.8GB |
| gpt-oss-20b | 21B | 16.6GB |
| gpt-oss-120b | 117B | 83.6GB |
Image models
| Model | Parameters | Memory |
|---|---|---|
| FLUX.1 schnell | 12B | 27.4GB |
| FLUX.1 dev | 12B | 27.4GB |
| Stable Diffusion XL Base 1.0 | 3B | 8GB |
| Stable Diffusion 3.5 Large | 8B | 19GB |
Video models
| Model | Parameters | Memory |
|---|---|---|
| HunyuanVideo | 13B | 60GB |
| Wan2.1 T2V 14B | 14B | 65.7GB |
| Wan2.2 T2V A14B | 27B | 80GB |
Audio models
| Model | Parameters | Memory |
|---|---|---|
| Whisper large-v3 | 1.55B | 3.6GB |