I’m looking for a tool or calculator that can estimate the minimum hardware needed to run a specific model locally.
For example, I want to know the cheapest setup that can realistically run Qwen 3.6 27B at decent speeds. Ideally something that can tell me:
- Required VRAM for different quantizations
- Whether it fits on a single GPU or needs multiple GPUs
- Expected tokens/sec
- RAM and CPU recommendations
- Power usage and rough total system cost
- Comparisons between setups like used 3090s vs newer cards
Does anything like this exist?
I know there are scattered benchmarks and Reddit posts, but I’m hoping there’s a more systematic tool or database people use when planning a local AI build.
[link] [comments]