Back to Models

DeepSeek V4 Pro

deepseekdeepseek/deepseek-v4-pro

How TheRouter serves this differently from the vendor

As the vendor operates it

DeepSeek serves V4-Pro first-party at api.deepseek.com and api.deepseek.com/anthropic, with OpenAI Chat Completions and Anthropic-compatible request formats, 1M context, 384K maximum output, thinking enabled by default at high effort, and no Responses API support for Pro on the cited pricing page.

On TheRouter

TheRouter exposes the same model through api.therouter.ai/v1/chat/completions as an OpenAI-compatible routed endpoint. Requests inherit TheRouter account routing, upstream selection, regional policy, tool-call normalisation, and TheRouter pricing rather than DeepSeek's first-party concurrency limits or first-party billing rates exactly.

API guide

Chat Completions

Standard OpenAI-compatible chat endpoint. Supports both non-thinking and thinking modes.

cURL
curl https://api.therouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $THEROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek/deepseek-v4-pro",
    "messages": [{"role":"user","content":"Explain quantum error correction in 3 sentences."}]
  }'

Thinking / Reasoning mode

Enable chain-of-thought via extra_body. Returns reasoning_content separately from final answer.

cURL
curl https://api.therouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $THEROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek/deepseek-v4-pro",
    "messages": [{"role":"user","content":"Solve: 9.11 vs 9.8 which is larger?"}],
    "extra_body": {"thinking": {"type": "enabled"}},
    "reasoning_effort": "max"
  }'

Tool calling

Native tool calling with full thinking-mode support. reasoning_content must be echoed back on subsequent turns when a tool call occurred.

cURL
curl https://api.therouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $THEROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek/deepseek-v4-pro",
    "messages": [{"role":"user","content":"What is the weather in SF?"}],
    "tools": [{"type":"function","function":{"name":"get_weather","parameters":{...}}}],
    "tool_choice": "auto",
    "extra_body": {"thinking": {"type": "enabled"}}
  }'

Recent coverage

Fact ledger β€” every claim on this page traces here
sourceURLretrieved
Release date (preview)api-docs.deepseek.com β†—2026-05-25verified
Architectureapi-docs.deepseek.com β†—2026-05-25verified
Context windowapi-docs.deepseek.com β†—2026-05-25verified
Max outputapi-docs.deepseek.com β†—2026-05-25verified
Reasoning modesapi-docs.deepseek.com β†—2026-05-25verified
License β€” codehuggingface.co β†—2026-05-25verified
License β€” weightshuggingface.co β†—2026-05-25verified
Open weightshuggingface.co β†—2026-05-25verified
Training tokensapi-docs.deepseek.com β†—2026-08-07unknown
Artificial Analysis Intelligence Indexapi-docs.deepseek.com β†—2026-08-07unknown
LiveCodeBench Pass@1api-docs.deepseek.com β†—2026-08-07unknown
MMLU-Proapi-docs.deepseek.com β†—2026-08-07unknown
GPQA Diamondapi-docs.deepseek.com β†—2026-08-07unknown
SWE-bench Verifiedapi-docs.deepseek.com β†—2026-08-07unknown
Terminal-Bench 2.0api-docs.deepseek.com β†—2026-08-07unknown
SimpleQA-Verifiedapi-docs.deepseek.com β†—2026-08-07unknown
BrowseCompapi-docs.deepseek.com β†—2026-08-07unknown
DeepSeek makes V4-Pro 75% discount permanentapi-docs.deepseek.com β†—2026-05-25verified
DeepSeek V4 series launches with 1M context and dual-tier MoE designapi-docs.deepseek.com β†—2026-05-25verified
When should I choose V4-Pro over V4-Flash?api-docs.deepseek.com β†—2026-08-07to verify
Does V4-Pro support the same thinking mode API as V4-Flash?api-docs.deepseek.com β†—2026-05-25to verify
Help & contact