Model Quant Measured Estimated RAM used App Source Date
Whisper Tiny Whisper Tiny FP16 12 tok/s 200 tok/s -94% 0.1 GB Ollama editorial 2026-03-20
Qwen 2.5 0.5B Qwen 2.5 0.5B Q4 8 tok/s 27 tok/s -70% 0.4 GB Ollama editorial 2026-03-20