| [ | |
| {"name":"Llama 3.2 3B Q4","paramsB":3,"minVramGB":4,"recommendedVramGB":6,"baseTokSec":42,"cmd":"ollama run llama3.2"}, | |
| {"name":"Qwen2.5 7B Q4","paramsB":7,"minVramGB":6,"recommendedVramGB":10,"baseTokSec":28,"cmd":"ollama run qwen2.5:7b"}, | |
| {"name":"Mistral 7B Instruct Q4","paramsB":7,"minVramGB":6,"recommendedVramGB":10,"baseTokSec":30,"cmd":"ollama run mistral"}, | |
| {"name":"Llama 3.1 8B Q4","paramsB":8,"minVramGB":8,"recommendedVramGB":12,"baseTokSec":24,"cmd":"ollama run llama3.1:8b"}, | |
| {"name":"Qwen2.5 14B Q4","paramsB":14,"minVramGB":12,"recommendedVramGB":16,"baseTokSec":15,"cmd":"ollama run qwen2.5:14b"}, | |
| {"name":"Llama 3.1 70B Q4 (multi-GPU/cloud)","paramsB":70,"minVramGB":40,"recommendedVramGB":80,"baseTokSec":3,"cmd":"vllm serve meta-llama/Llama-3.1-70B-Instruct"} | |
| ] | |