deepseek-v4-pro

Name: deepseek-v4-pro
Author: Vercel AI Gateway

mit

DeepSeek-V4-Pro hybrid model with both non-thinking and thinking (default) modes.

Context

Max output

384K

Input price

$0.43/1M tokens

Output price

$0.87/1M tokens

Capabilities

vision

tool call

structured output

reasoning

json mode

streaming

fine tuning

batch

Details

Model IDdeepseek/deepseek-v4-pro

ProviderVercel AI Gateway

CreatorDeepSeek

Familydeepseek

Licensemit

Parameters—

Statusactive

Input modalitiestext

Output modalitiestext

Architecture—

Knowledge cutoff—

Training data cutoff—

Release date2026-08-13

Deprecation date—

Typechat

Reasoning tokensYes

Max input—

Open weightYes

Sourceofficial

Last updated2026-05-01

Tools

Function Callingfunction_calling

Call external functions and APIs

Pricing

Input

$0.43

↓75%$1.74

Output

$0.87

↓75%$3.48

Cache write

—

Cache read

—

Batch in

—

Batch out

—

Price history · $/1M input

$1.74

$0.43

Apr 25May 2

Date	Input	Output	Change
Apr 25, 2026	$1.74	$3.48	—
May 2, 2026bc0fb7d	$0.43	$0.87	↓75%

Input is ↓75% since Apr 25, 2026, from $1.74 to $0.43 per 1M tokens.

Price across providers

Provider	Input	Output	Since launch
Vercel AI Gatewaycheapest	$0.43	$0.87	↓75%
OpenRouter	$1.17	$2.34	↓33%
Novita AI	$1.6	$3.2	↓5%

Family Comparison: deepseek

Model	Caps	Perf	Speed	Context	Max out	Pricing
deepseek-v4-flash-0731	RVTS	—	—	1M	—	$0.08$0.15
deepseek-v4-flash	RVTS	—	—	1M	384K	$0.09$0.18
deepseek-v3.1	RVTS	—	—	164K	—	$0.25$0.95
deepseek-v3.2-thinking	RVTS	—	—	164K	—	$0.26$0.38
deepseek-v3.1-terminus	RVTS	—	—	131K	—	$0.27$1
deepseek-v3	RVTS	—	—	164K	—	$0.27$1.12
deepseek-v3.2	RVTS	—	—	164K	—	$0.28$0.42
deepseek-v4-pro-0813	RVTS	—	—	1M	—	$0.43$0.87
deepseek-v4-pro	RVTS	—	—	1M	384K	$0.43$0.87

API

GET/v1/models/vercel/deepseek/deepseek-v4-pro