DeepSeek-V3

Name: DeepSeek-V3
Author: DeepInfra

deepseek

DeepSeek-V3, a strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token. To achieve efficient inference and cost-effective training, DeepSeek-V3 adopts Multi-head Latent Attention (MLA) and DeepSeekMoE architectures, which were thoroughly validated in DeepSeek-V2.

Context

164K

Input price

$0.32/1M tokens

Output price

$0.89/1M tokens

Capabilities

Show unsupported (3)

Reasoning & Interaction

Tool Calling

Output Control

Structured Output

JSON Mode

Infrastructure

Streaming

Fine-tuning

Details

Model IDdeepseek-ai/DeepSeek-V3

ProviderDeepInfra

CreatorDeepSeek

Familydeepseek-v3

Licensedeepseek

Statusactive

Typechat

Input modalitiestext

Output modalitiestext

Open weightYes

Parameters—

Architecture—

Context window164K tokens

Max input—

Max output—

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.deepinfra.com/v1/openai/v1/chat/completions

Pricing

Input

$0.32

Output

$0.89

Cache write

—

Cache read

—

Batch in

—

Batch out

—

Family Comparison: deepseek-v3

Model	Caps	Perf	Speed	Context	Max out	Pricing
DeepSeek-V3-0324	TJJMSF	—	—	164K	—	$0.24$0.9
DeepSeek-V3.1	TJJMSF	—	—	164K	—	$0.25$0.95
DeepSeek-V3.2	TJJMSF	—	—	164K	—	$0.26$0.38
DeepSeek-V3.1-Terminus	TJJMSF	—	—	164K	—	$0.27$0.95
DeepSeek-V3	TJJMSF	—	—	164K	—	$0.32$0.89

API

GET/v1/models/deepinfra/deepseek-ai/DeepSeek-V3