Directory

Local inference hardware

This directory tracks 10 GPUs, accelerators and SoCs used for local model inference, with VRAM, memory bandwidth, power draw and list price. Specifications are curated by hand from vendor datasheets and carry the date a person last checked them; fabrication details are read from Wikipedia. Every figure shows which of the two it is.
Machines
10
VRAM catalogued
792
With a listed price
5/10
Last sync
just nowsucceeded
RTX A6000VRAM, GB48editorial · 29d agoInstinct MI300XVRAM, GB192editorial · 29d agoRadeon RX 7900 XTXVRAM, GB24editorial · 29d agoApple M4 ProVRAM, GB64editorial · 29d agoApple M2 UltraVRAM, GB192editorial · 29d agoApple M3 MaxVRAM, GB128editorial · 29d agoH100 PCIeVRAM, GB80editorial · 29d agoGeForce RTX 3090VRAM, GB24editorial · 29d agoGeForce RTX 4080 SuperVRAM, GB16editorial · 29d agoGeForce RTX 4090VRAM, GB24editorial · 29d ago

No hiring column here. The job-board classifier extracts framework names, so the join can only speak about frameworks — for these entities we have not measured demand, rather than measured it to be zero.

3 GPUs, accelerators and SoCs, sorted by vram. Each row shows the last time the ranked figure was observed.
Select to compareMachineVRAM, GBGB/sTDP, WMSRPClassChecked
Apple M2 UltraApple192800295not listedsocCurrent29d ago
Apple M3 MaxApple12840078not listedsocCurrent29d ago
Apple M4 ProApple6427346not listedsocCurrent29d ago

1 of 3 machines selected — pick one more to compare

CompareClear

Source: Wikipedia infobox. Every value carries the moment it was observed; figures that failed validation are withheld rather than shown. SLA: current within 720h, ageing to 2160h, stale beyond 4320h.

indexable · canonical /data/hardware?vendor=apple

Also tracked

Other datasets

Same engine, same provenance rules — every figure carries the source that reported it and when.

Smoke dataset

A fixture dataset that exercises the ingestion, validation and freshness machinery.

4 tracked

Agent engineering frameworks

Open-source frameworks for building, orchestrating and evaluating LLM agents, tracked across GitHub, npm and PyPI.

18 tracked

Language and embedding models

Language, embedding and speech models tracked across Hugging Face Hub download activity and OpenRouter context windows and pricing.

18 tracked