Observed arrival · 2026-08-28
The Hitchhiker's Guide to Open-Weight Models
A humorous reference guide comparing downloadable AI models by capability, benchmark results, licensing, and the hardware needed to run them.
Field notes
The guide organizes models by the hardware sacrifice required to run them, from Frontier systems to an intentionally comic “Towel” tier. Entries expose practical comparison data including parameter counts, active parameters, context length, licensing, benchmark scores, and estimated VRAM. For example, the visible Kimi K3 entry lists 2.8 trillion total parameters, roughly 50 billion active parameters, a 1,048,576-token context, and an estimated 5.6 TB of FP16 VRAM.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue