Chinese AI Tools
ZH

DeepSeek

DeepSeek V4 API

The English DeepSeek docs list deepseek-v4-pro and deepseek-v4-flash as the current models. Both support thinking and non-thinking modes, JSON output, tool calls and chat prefix completion; FIM completion is available in non-thinking mode. The docs also state that deepseek-chat and deepseek-reasoner will be deprecated on 2026-07-24 and currently map to V4-Flash modes for compatibility.

Globally availableFull English UIPublic APIPaidTrusted

Quick answers

At a glance

Overview
DeepSeek's current V4 API line, covering V4-Pro and V4-Flash with 1M context, thinking mode and tool calls.
Best fit
Developers migrating DeepSeek integrations to current V4 model names and long-context reasoning workflows.
Trust
3/3 sources verified, recently checked · 2026-05-17
Coverage
100/100

Editorial verdict

Best for

Developers migrating DeepSeek integrations to current V4 model names and long-context reasoning workflows.

Avoid if

Avoid starting new projects on deepseek-chat or deepseek-reasoner aliases.

Why it matters

V4 is the current API line, with separate migration and pricing implications from the broader DeepSeek brand.

Pricing

V4-Flash starts at documented per-1M-token pricing; V4-Pro has a documented temporary discount through 2026-05-31 15:59 UTC

Payment

Topped-up balance, Granted balance, Platform billing

Commercial use

Commercial use should follow the current product, API, model license and billing terms.

Privacy

Review prompt, file, media upload, retention and training-use terms before sensitive workloads.

Use-case fit

V4 model integration

Strong

Use deepseek-v4-pro for agent/coding tasks and deepseek-v4-flash for faster or lower-cost workloads.

Thinking mode control

Strong

Use the documented thinking toggle and effort controls for reasoning-heavy requests.

Global user checklist

RegistrationConfirmedRequires a DeepSeek Platform API key.
English UIConfirmedQuick start, pricing, guides and change log are English-facing.
API and docsConfirmedDocs cover OpenAI and Anthropic formats plus V4 feature details.
International paymentPartialPricing is clear, but account-specific top-up/payment routes still need verification.
Coverage · 100/100

Model names, quotas, release status, regional access and commercial terms can change quickly; recheck official sources before procurement or production use.

Pros

  • - 1M context and 384K maximum output are documented
  • - Thinking and non-thinking modes share one current model line
  • - Tool calls, JSON output and context caching are documented

Cons

  • - Legacy aliases require migration before 2026-07-24
  • - V4-Pro pricing includes a time-bounded discount

Decision paths

qwen

kimi-k2-api

zhipu-glm

Sources

DeepSeek models and pricing

pricing · en · verified 2026-05-17

Lists V4-Pro/V4-Flash details, pricing and deprecation note.

DeepSeek thinking mode

docs · en · verified 2026-05-17

Documents thinking toggle, effort control and reasoning_content behavior.

DeepSeek change log

docs · en · verified 2026-05-17

Confirms the 2026-04-24 V4 release and 2026-07-24 alias deprecation.

Reviews