LLMilestonesA milestone timeline of AI and large language models

Act V · Reasoning and agents

PaperDeepSeek

DeepSeek-R1 published in Nature

The R1 paper was peer-reviewed and published in Nature.

$294Kto train R1 (on top of about $6M for the base model)

Why it matters

Believed to be the first major large model to pass peer review; it disclosed that R1's reasoning training cost only about $294,000.

Built on

DeepSeek-R1

Source: nature.com ↗

Official page not opened for verification; URL taken from citations in multiple reports.

See it on the 3D timeline →