viraldi.
Video details
How it works
Back to discovery
13:35
Technology & Computing / VIDEO

I Tested Every Local AI Model So You Don't Have To

Kai· 58.4K subscribers

29d ago · English audio

The best local coding model is the same on a 12GB card and a 48GB one, and every GPU tier list misses it. That model is Qwen 3.8 27B. On a bigger card it doesn't get smarter; you get more bits and a longer context. The number most tier lists skip is the KV cache, the running notes a model keeps on every file your coding agent reads. In Qwen 3.8 27B only 16 of its 64 layers keep notes that grow, so a full 128K…

Watch on YouTube
4.6×Score at discovery
75.9KViews
16.4KMedian · 50 videos

Outlier score

This score was measured on Sun, 25 Oct 2026 20:21:48 GMT. Current view counts may differ.

75.9K views ÷ a 16.4K baseline = 4.6×. Compared with 50 earlier videos from the same channel, matching the format recorded at scoring time. Older-upload baseline: reference videos had at least as long to gain views. This comparison is not age-matched.

YouTube snapshot: Sun, 25 Oct 2026 22:36:18 GMT. Counters may have changed.

Read the methodology

Generative AI & Agents

Channel-balanced percentile: 96

Comparison details

35 videos from 29 other channels, with each channel weighted equally. Same format and language, with 10,000–99,999 subscribers. Observed at 703 hours after publication, within ±48 hours.

This describes the collected sample, not all of YouTube. Updated Sun, 25 Oct 2026 22:46:31 GMT.

Niches & tags

Transcripts and detailed analytics are available for recent outlier Shorts up to three minutes. Research tools

Breakouts from other channels