Observed arrival · 2026-09-10

VRAM Spec: A Field Guide to Running AI Models on Your GPU

vramspec.com Observed source

Editorial interest 78/100 Selection signal · not a rating of the site

A browser-based reference for estimating the memory requirements of LLMs, image models, and video models.

Landing page captured for the 2026-09-10 issue.

Field notes

The homepage organizes 20 entries into 15 LLM models and five image or video models, pairing each with a minimum VRAM estimate and example hardware. The scale is easy to grasp: DeepSeek-R1-Distill-Qwen-1.5B is listed at 2.5 GB, while the 671B DeepSeek-R1 entry reaches 380 GB and names eight H100s or a Mac Studio M2 Ultra with CPU offload. Wan 2.1 is split into 1.3B and 14B variants.

Observed signals

Read the marks

Editorial observations of this landing page, not a rating.

○OpenPublic substance visible

✦PrettyNotable craft visible

●ProPolished or operationally mature

◎NicheUnusually specific use

One card from the complete issue

All Approaches to Piccadilly

405,063 arrived 1,000 judged 996 catalogued Enter the complete issue