VRAMGlass

GPU comparison, VRAM planning and price tracking for local LLMs

FreeAI Developer Tools Large Language Models (LLMs)Open Source AI Models

Are you the maker of VRAMGlass?Advertise

What is VRAMGlass?

VRAMGlass is a free web tool for people choosing hardware to run large language models locally. It brings GPU specs, model memory estimates and market price references together, so you can plan the hardware you actually need. The site is available in English, Simplified Chinese, Traditional Chinese and Japanese.

What can VRAMGlass do?

GPU comparison
Compare GPUs side by side by VRAM capacity, memory bandwidth, power and software ecosystem.
Model memory and fit estimates
Explore open-weight model families and their quantized weight sizes, then use documented memory and fit estimates to see what a model needs.
Multi-market GPU price references
Separate references for US used listings on eBay, Chinese used listings on Xianyu and Japanese new offers on Amazon.co.jp, with price histories, sample counts, observation dates and stale-price labels.
RAM price tracking
Desktop and server RAM price references sit alongside the GPU data.
Practical guides
Guides cover VRAM, system RAM, context and KV cache, quantization and offloading.
Clear data limits
Prices are asking-price references, not completed sales. Fit and performance estimates are planning aids rather than real-world benchmarks.

Who is VRAMGlass for?

Local LLM builders: Pick a model and quantization, estimate the memory requirement, then compare GPUs that fit.
Used GPU buyers: Review eBay, Xianyu or Amazon Japan price samples and the methodology behind them before buying.
Home lab and server planners: Plan system RAM for offloading and check desktop and server RAM prices.