DeepSeek V4 Flash

reasoning

DeepSeek V4 Flash is an open-weight reasoning model with a 262k context window, hosted by Reka. It is served as an OpenAI-compatible passthrough with native tool calling, structured outputs, logprobs, and streaming.

Context
262K
Max output
131K
Input price
$0.11/1M tokens
Output price
$0.66/1M tokens

Capabilities

Reasoning & Interaction
Reasoning
Tool Calling
Output Control
Structured Output
JSON Mode
Infrastructure
Streaming

Details

Model IDdeepseek4-flash
Provider Reka AI
Creator DeepSeek
Familydeepseek
License—
Statusactive
Typereasoning
Input modalitiestext
Output modalitiestext
Open weight—
Parameters—
Architecture—
Context window262K tokens
Max input—
Max output131K tokens
Reasoning tokens—
Knowledge cutoff—
Release date—
Sourceofficial

Tools

Function Callingfunction_calling
Call external functions and APIs

Endpoints

Chat CompletionsPOST
Generate chat responses with messageshttps://api.reka.ai/v1/chat/completions

Pricing

Input
$0.11
Output
$0.66
Cache write
—
Cache read
$0.01
Batch in
—
Batch out
—

Family Comparison: deepseek

ModelContextMax outPricing
DeepSeek V4 Flash262K131K$0.11$0.66
DeepSeek V4.1 Flash 10171.0M393K$0.15$0.6
DeepSeek V4.1 Flash1.0M393K$0.3$1.2

API

GET/v1/models/reka/deepseek4-flash