Unverified50% confidenceFactExact time

The 124M scale is the training capacity upper limit that a single consumer-grade GPU (such as an RTX 3090/4090) can handle.

1

Sources

50%

Confidence

Medium-term (~90 days)

Relevance

10/11/2026

First Seen

Valid until: 1/9/2027

Sources

Building Levy-1 from Scratch: A 124M Parameter Small Model and What to Improve Next

Related Entities

RTX 3090product RTX 4090product

Related Claims

VerifiedThe consumer-grade GPU RTX 4090 has a VRAM ceiling of 24 GB.72% similar Unverified2B-parameter-scale models can run on a single consumer-grade GPU (e.g., RTX 3090/4090 with 24 GB VRAM) or even in some CPU inference scenarios.71% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/996335

API

curl https://kongchang.com/api/v1/knowledge/claims/996335

MCP

get_claim(id=996335)