Model family timeline
Last updated May 28, 2025
DeepSeek R1 model releases
A source-backed timeline for the DeepSeek R1 model family, collecting release dates, labs, access details, context windows, and major lifecycle changes.
3
Models
1
Labs
2
Open
3
Recent
3 models
DeepSeek-R1-0528
AvailableDeepSeekFrontierOpen source
Major R1 reasoning update with stronger math, programming, general logic, function calling, and reduced hallucinations.
Type
MoE
Params
671B
Context
128K
Top eval
—
Released
May 28, 2025
1.1y ago
MoE671B128K ctxMay 28, 2025
DeepSeek-R1
AvailableDeepSeekFrontierOpen source
Breakout open reasoning model trained with large-scale reinforcement learning and released with weights under MIT.
Type
MoE
Params
671B
Context
128K
Top eval
—
Released
Jan 20, 2025
1.4y ago
MoE671B128K ctxJan 20, 2025
DeepSeek-R1-Lite-Preview
RetiredDeepSeekProprietary
Reasoning-preview model exposed in DeepSeek Chat ahead of the open DeepSeek-R1 release.
Type
—
Params
Undisc.
Context
—
Top eval
—
Released
Nov 20, 2024
1.6y ago
—Undisc.— ctxNov 20, 2024