You could have both worlds: an LLM model by IBM https://huggingface.co/ibm-granite/granite-4.0-h-small

It wasn't very promising when it came to benchmarks though, go figure: https://artificialanalysis.ai/leaderboards/models