Model index
DeepSeek·DeepSeek V / R

DeepSeek-V4-Flash-Preview

The efficient V4 preview with 284B total, 13B active parameters, and a 1M context window.

LegacyLanguageReasoningCodingAgentOpen weights
MODEL NUMBER065
Release date
2026-04-24
Organization
DeepSeek
Family
DeepSeek V / R
01

Specifications

Only published specifications are shown.

Total parameters
284B
Active parameters
13B
Architecture
MoE
Context
1M
Input modalities
Text
Open weights
Yes
License
DeepSeek Model License
Status
Legacy
02

Overview

Overview

The efficient V4 preview with 284B total, 13B active parameters, and a 1M context window.

03

Changes from predecessor

Compared with DeepSeek-V4-Pro-Preview

PreviousDeepSeek-V4-Pro-Preview1.6T · 1M ctx
CurrentDeepSeek-V4-Flash-Preview284B · 1M ctx
New capabilities Hybrid reasoning Long context
04

Architecture & capabilities

Attention

DSA

Key capabilities

Hybrid reasoningAgentCodingLong context
05

Release history

Preview

DeepSeek-V4-Flash Preview launched with V4-Pro Preview and open weights.

View source
06

Related models

Select a card to view the related model.

07

Sources