You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Looking for real-world experience, not spec sheets.
If you've run models on-device, please share:
Device (e.g. iPhone 15 Pro, Pixel 9)
Model + quant (e.g. Qwen2.5 1.5B Q4_K_M)
Tokens/sec you actually see
Quality verdict for chat / summarization / drafting
My starting point: on Apple-Intelligence devices the built-in ~3B system model is hard to beat (zero download, NPU-accelerated). For everything else, 1–1.5B Q4 models feel like the sweet spot between speed and usefulness. What's your data?
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Looking for real-world experience, not spec sheets.
If you've run models on-device, please share:
My starting point: on Apple-Intelligence devices the built-in ~3B system model is hard to beat (zero download, NPU-accelerated). For everything else, 1–1.5B Q4 models feel like the sweet spot between speed and usefulness. What's your data?
All reactions