Explore Local Models
Hardware-aware scatter plot and table of every local LLM. Click any model to compare.
X axis · Y axis
- VRAM required
- Disk size
- Model size
- Est. tok/s
- Quality index
- Context
Filters
Best value · Fastest · Fits my hardware
tok/s are bandwidth-based estimates, not benchmarks. Context 8,192 tokens, KV cache f16.
Quality index from Artificial Analysis, where available.
Click any point or row to open it in Compare.
More ways to choose a model
Models section
Tools