# Gemma4-12B Cross-Platform Residency and Benchmarking > Source-anchored research for Gemma4-12B Cross-Platform Residency and Benchmarking. Run status: COMPLETED. Parent: [Local Model Performance Evaluation: Ollama, MLX, and Hardware Architectures](https://llms-explorer.com/tree/local-model-performance-evaluation/) · 3 facets · 4 facts · page: https://llms-explorer.com/tree/gemma4-12b-cross-platform-residency-and-benchmarking/ ## Measurements and reference values - Gemma4-12B quantized at 4-bit occupies ~7.8 GB VRAM, leaving 8.2 GB headroom in a 16GB VRAM pool for up to 32k token KV cache and activation workspaces. (high confidence) — [source](https://github.com/google/gemma_pytorch) - When running completely resident in RTX 5080 GDDR7 VRAM (~1000 GB/s), Gemma4-12B generates ~85-92 tokens per second, roughly double the ~42-45 tokens per second achieved on Apple Silicon M5 Max (~400 GB/s). (high confidence) — [source](https://github.com/tinygrad/tinygrad) ## Comparisons and alternatives - Serving Gemma4-12B via Ollama Metal introduces a 10-15% throughput penalty compared to native MLX-LM due to runtime translation and generic Metal kernel dispatch in ggml. (high confidence) — [source](https://github.com/ml-explore/mlx-examples) ## Facts and statements - Pre-allocation of static tensor workspaces in Tinygrad or TensorRT avoids CUDA and Metal memory fragmentation during extended generation passes. (high confidence) — [source](https://github.com/tinygrad/tinygrad)