Resource detail

DeepSeek-V2

The MoE model that drew attention with exceptionally low inference cost.

DomesticFreeOpen sourceChinese support

Visit official siteMore like this

At a glance

deepseek-v2

Pricing

Free

Access

No login

Platforms

Web

Chinese support

Yes

Last checked

2026-08-29

Popularity

83/100

Model background

Open weightsRetired archive

DeepSeek-V2's MLA attention and low-cost serving reset price expectations and laid engineering foundations for V3 and R1.

Why it mattered

DeepSeek-V2's MLA attention and low-cost serving reset price expectations and laid engineering foundations for V3 and R1.

Official linkhttps://www.deepseek.com/

Model capabilities

Multimodal -Reasoning ✓Coding ✓Tool use ✓Long context ✓

Context length

128K tokens

Provider

deepseek

Typical use

低成本 APIResearch

Similar tools and resources

Selected by use case

DeepSeek-V3

DeepSeek's open general model for cost-effective chat and coding.

DomesticFreemiumOpen source+2

Updated 2026-08-29

Visit official site

Mixtral 8x7B

An open MoE model that made sparse, efficient inference practical.

GlobalFreeOpen source+2

Updated 2026-08-29

Visit official site

GPT-5

OpenAI's flagship model with adaptive reasoning and fast responses.

OverseasPaidFlagship+2

Updated 2026-08-29

Visit official site

GPT-4o

OpenAI's multimodal model for general Q&A, vision, and tool use.

OverseasPaidMultimodal+2

Updated 2026-08-29

Visit official site

OpenAI o3

OpenAI's deep-reasoning model for math, science, and complex planning.

OverseasPaidReasoning+2

Updated 2026-08-29

Visit official site

Claude Sonnet 4

Anthropic's high-performance model for coding, long documents, and complex reasoning.

OverseasPaidLong context+2

Updated 2026-08-29

Visit official site