gemma-4-12b-it vs Qwen3.5-9B on shared benchmarks: Qwen is overall winner beating gemma in 5/8 benchmarks despite a smaller footprint
A comparison of the Gemma-4-12B-it and Qwen3.5-9B large language models indicates that Qwen generally outperforms Gemma on a per-gigabyte basis. The Qwen model achieved better results in 5 out of 8 benchmarks, despite having a smaller parameter footprint. While Gemma-4-12B-it may show slightly superior coding capabilities, specialized fine-tunes of Qwen are available for such tasks. AI
IMPACT Qwen3.5-9B demonstrates competitive performance against larger models, potentially influencing choices for efficient local deployments.