Feel free to contribute.

gemma4:e4b (9.6GB) - Ollama

Practically useless for complex planning tasks it fails expanding the code base and working alone one tasks

gemma-4-26b-a4b - LM Studio

It is working much better - even if it is the same model; but much larger - expanding code and planing properly.

LM Studio is unstable v0.4.10

gemma4:26b / gemma4:26b-a4b-it-q4_K_M - Ollama

As good as gemma-4-26b-a4b - LM Studio, but faster and stable.

Qwen 3.5 - LM Studio

Currently not working properly due to an LM Studio Bug

https://github.com/lmstudio-ai/lmstudio-bug-tracker/issues/1592