Recipes describe a configuration that worked (or didn't work).
+ Add a recipeModel
Harness
Hardware
Interface
Use cases
lmstudio-community/gemma-4-12B-it-QAT-GGUF Q4_0
LM Studio
Great for vision, strong improvement over the non-QAT version
chat vision
| Hardware | Runs | Works | |
|---|---|---|---|
| M4 Pro | 1x |
lmstudio-community/gemma-4-12B-it-GGUF Q4_K_M
LM Studio
Great at image understanding but can hallucinate
chat reasoning vision
| Hardware | Runs | Works | |
|---|---|---|---|
| M4 Pro | 1x |
bartowski/Qwen_Qwen3.5-2B-GGUF Q4_K_M
llama-server
Good tool-calling model for limited compute settings
chat tool-use
| Hardware | Runs | Works | |
|---|---|---|---|
| Cortex-A76 | 1x |
lmstudio-community/Qwen3.5-9B-GGUF Q4_K_M
LM Studio Pi
Decent coding model, can run on a 24GB macbook
coding
| Hardware | Runs | Works | |
|---|---|---|---|
| M4 Pro | 1x |