DeepSeek: DeepSeek V4 Flash 0731

deepseek/deepseek-v4-flash-0731

25th smartest of 183Cheaper than 78% of paidReasoningToolsJSON

Use via OpenRouter ↗

Intelligence

49.9

25th of 183

Design Elo

1246

Speed

—

tokens/sec

Latency

—

first token

Input price

$0.140

72nd cheapest

Context

1.0M

384K max out

How it compares

Smarter than86%

of all ranked models

Cheaper than78%

of all ranked models

Overview

DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows.

Benchmarks

independent · Artificial Analysis & Design Arena

Artificial Analysis95th percentile

Intelligence Index

49.9

Coding Index

69.1

GPQA Diamond

91%

Humanity's Last Exam

37%

SciCode

50%

Design Arena · Elo7,924 tournaments

1246

Website

1232

svg

1200

Data Viz

1156

Providers & pricing (1)

Provider	In $/M	Out $/M	Context	Uptime
DeepSeekfp8cache	$0.140	$0.280	1.0M	100%

Specifications

Context window1.0M

Max output384K

Knowledge cutoff—

Input modalitiestext

Output modalitiestext

Prompt cachingSupported

Cache read price$0.0028/M

ModeratedNo

DeepSeek V4 Flash 0731 FAQ

How much does DeepSeek V4 Flash 0731 cost?

DeepSeek V4 Flash 0731 costs $0.140 per million input tokens and $0.280 per million output tokens via OpenRouter, making it 72nd cheapest of 333 paid models.

How smart is DeepSeek V4 Flash 0731?

DeepSeek V4 Flash 0731 scores 49.9 on the Artificial Analysis Intelligence Index, ranking 25th of 183 benchmarked models, with a GPQA Diamond score of 91%.