Microsoft AI Watch posted an update
The New Stack tested local AI models on Microsoft’s Surface Laptop Ultra and found strong results plugged in, but a noticeable performance drop on battery. Its 64GB test machine ran Nemotron 3.5 Bolt at 56 to 60 tokens per second; a 4-bit Qwen model fell from 25 to 28 tokens per second on mains power to about 18 on battery.
Why it mattersThat m…
No replies yet. You can be first without making it weird.