Production routeThis rate reflects platform-side availability — only confirmed server errors (HTTP 500 / empty response) count as failures. User-side issues (content moderation, invalid params, cancellation) plus rate limits, timeouts and auth errors are excluded. Before real traffic arrives, empty buckets may display as available.Live
Live
Context
1M context · 384K max output
Best For
Deeper reasoning, complex coding, and agent runs where Flash falls short
Input
Text
Output
Text · JSON (structured output)
DeepSeek plans to switch deepseek-v4-pro to V4.1 Flash on its direct API on September 14, 2026; on EvoLink, deepseek-v4-pro is not affected and still serves V4 Pro. See the V4.1 Flash migration guide.
An evaluation route for coding, reasoning, and long-context analysis, available on Chat Completions, Messages, and Responses — use whichever protocol your stack already speaks.
DeepSeek V4 Pro
DeepSeek reasoning and tool-use model
Selected
Model ID
deepseek-v4-pro
Best for
High-volume coding, reasoning, long-document analysis, and agent workloads that need a low-cost route with cached-input pricing.
Input
$1.324 / 1M
90 cr / 1M
Cache hit
$0.045 / 1M
3 cr / 1M
Output
$3.971 / 1M
270 cr / 1M
All rates are per 1M tokens, shown in USD and credits, and reflect your account's current pricing.
DeepSeek V4 Pro pricing
Estimate what one DeepSeek V4 Pro request costs before you integrate. The calculator uses your account's current rates, with official pricing as a reference.
Request calculator
Enter the token mix for one request and the number of successful tool calls.
Only successful server-side calls are billed per call; failed attempts have no tool fee, but tokens still apply.
Web search$0.010/ call·0.68 cr / call
What is the DeepSeek V4 Pro API?
DeepSeek V4 Pro is DeepSeek’s strongest tier: a 1.6-trillion-parameter Mixture-of-Experts design with 49B active parameters per token, a 1M-token context window, 384K maximum output, and thinking mode enabled by default. On EvoLink it is available on both /v1/chat/completions (OpenAI style) and /v1/messages (Anthropic style) using the model ID deepseek-v4-pro. Review pricing boundaries, supported workflows, and model configuration in the Pricing and API sections below. DeepSeek plans to forward deepseek-v4-pro on its direct API to V4.1 Flash from September 14, 2026 at 12:00 Beijing time (04:00 UTC); on EvoLink, deepseek-v4-pro is not affected and continues to serve V4 Pro.
DeepSeek V4 Pro is served on EvoLink under the model ID deepseek-v4-pro through Chat Completions · Anthropic Messages · Responses, with the same API key and balance you use for every other model. It offers a 1M context window and up to 384K output tokens, plus reasoning, long-context coding, prompt caching.
DeepSeek V4 Pro specs and capabilities
Numbers come from the EvoLink route configuration; capabilities are what the API exposes today.
Context window
1M tokens
Max output
384K tokens
Input
Text
Output
Text · JSON (structured output)
Reasoning
Thinking mode (on by default)
Prompt caching
Automatic cache reads at a lower rate
Server-side tools
Web search, billed per successful call
Protocols
Chat Completions · Anthropic Messages · Responses
Model ID
deepseek-v4-pro
Requests on Chat Completions, Messages, and Responses must use the exact model ID shown here.
What is the DeepSeek V4 Pro API best suited for?
DeepSeek V4 Pro combines a 1M-token context window, configurable reasoning, and tool calling for work that must retain evidence, call external systems, and return verifiable results. The model is text-only — it does not accept image input. These use cases explain where it may add value and what to confirm before integration.
Repository-scale coding and code review
Keep relevant source files, issues, test results, and previous changes in one task for cross-file debugging, implementation planning, and review. Before rollout, use fixed repository tasks to check test pass rate, incomplete steps, structured results, and required human edits instead of judging coding ability from a single demo.
Long documents and multi-source analysis
The 1M-token context can hold reports, contracts, knowledge-base passages, conversation history, and retrieved sources together. More context does not automatically improve an answer, so verify that key evidence is preserved and use the live rates and cache-hit mechanics in Pricing to calculate cost per acceptable result.
Long-horizon agent and terminal workflows
The 0813 GA build sharply improved agentic coding and terminal-task performance, which is where community reports (a demand signal, not a verified benchmark) most often place V4 Pro’s value against closed models. Pair it with your own retrieval and execution tools through function calling, and validate multi-step task completion rates on your real workloads before scaling.
Structured output and agent orchestration
Text inputs can feed JSON Schema, function calling, and multi-step agent flows for extraction, review, and downstream automation. Before production use, test schema validity, function arguments, streaming completion events, and safe recovery after a tool failure.
What changes when you access DeepSeek V4 Pro through different platforms?
Even with the same model, platforms can differ in context configuration, tool support, usage reporting, and billing. EvoLink puts model configuration and usage behind one API so later model changes remain straightforward.
Use the correct API model ID
The page URL and the API request model ID are the same string: deepseek-v4-pro. Existing Chat Completions or Responses applications can use the correct ID in model configuration, with exact fields available in the API section.
Follow the active API route configuration
DeepSeek documents a 1M-token model window, but platforms may expose different context settings, tools, and rate limits. When using EvoLink, rely on the model configuration, available features, and actual usage shown for the current route.
Choose Chat or Responses by workflow
Start with Chat Completions for standard chat, streaming, and client-side functions. Choose Responses — added with the 0813 build — for multi-step agent workflows and its documented server tools: function calling, web search, and apply_patch. Note that code interpreter is ignored on this route. This preserves a familiar OpenAI-style integration without adding complexity the workload does not need.
Keep model choice open in one gateway
Use one EvoLink account, balance, and API pattern for Grok, GPT, Claude, and Kimi. Keeping model selection in configuration lets teams route by task quality, cost, and availability without rebuilding application code for every provider.
How can you control DeepSeek V4 Pro API cost more accurately?
The existing Pricing section shows current token rates (input, cache hit, output). In practice, caching, context management, reasoning settings, and model routing help reduce unnecessary usage and connect spend to completed business tasks.
Make repeated context cache-friendly
Stable system prompts, tool schemas, and shared context are easier to reuse through caching. Configure the supported cache or conversation identifier for the selected protocol and inspect cached tokens in usage to confirm that repeated requests receive the expected benefit.
Send only the context the task needs
A 1M-token window is useful for large repositories and documents, but it does not need to be filled on every request. Selecting only relevant files, messages, and retrieved passages reduces input cost and helps the model focus on the evidence that matters.
Tune reasoning, output, and tool calls by task
Start simple work with lower reasoning effort and shorter output, then increase the reasoning budget for harder analysis. Research and agent workflows should also track tool-call count to avoid repeated searches, executions, or unproductive loops.
Compare models by total task cost
Evaluate tokens, cached input, server tools, and required retries within the same completed task. EvoLink centralizes model and usage information so teams can compare the total cost of completing equivalent work across DeepSeek V4 Pro and other routes.
Two ways to use DeepSeek V4 Pro: EvoLink API or Agent
Use the EvoLink API for product backends and batch jobs, or call DeepSeek V4 Pro from Codex, Claude, or Gemini for coding and analysis workflows. Both paths share the same EvoLink API key, balance, model ID, and request history.
Option 1
Integrate with the EvoLink API
Best for: product backends, batch jobs, automated pipelines
Send OpenAI-compatible Chat Completions (or Anthropic Messages) requests to EvoLink and control the model ID, system prompt, output budget, tools, and structured output.
1Create an EvoLink API key in the console
2Point your OpenAI or Anthropic SDK at the EvoLink base URL and select the model ID shown above
3Send one representative request and read the usage field for input, cached, and output tokens
4Set max_tokens and retries per task; keep tool-call IDs and results across turns
Best for: coding, review, and analysis tasks in Codex, Claude, and Gemini
Give the Agent the task, the inputs to include, and the acceptance criteria. It assembles the request, calls DeepSeek V4 Pro through EvoLink, and returns the answer with token usage.
1Set EVOLINK_API_KEY in your local environment; never put it in code or a prompt
2Describe the task, the inputs to include, and the expected output format
3Ask the Agent to call DeepSeek V4 Pro through EvoLink and show the request before sending
4Let the Agent report the answer, token usage, and any error body
DeepSeek V4 Pro API code example and error handling
This example shows the shortest runnable request: an OpenAI-compatible Chat Completions call with a system prompt, a user message, and an output budget. Open the API tab for the complete parameter and response reference.
curl -X POST https://api.evolink.ai/v1/chat/completions \
-H "Authorization: Bearer $EVOLINK_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-v4-pro",
"messages": [
{ "role": "system", "content": "You are a senior engineer reviewing a pull request." },
{ "role": "user", "content": "Review this diff and list every risky change with file and line:\n<diff>" }
],
"max_tokens": 4096,
"temperature": 0.2
}'
# Anthropic Messages (/v1/messages) and Responses (/v1/responses) accept the
# same model ID.
# The response includes choices[0].message and a usage object (prompt,
# cached, and completion tokens).
Invalid request or unsupported parameter
Check the model ID, the messages array, and parameter ranges against the API reference; remove fields this route does not support.
Authentication or balance issue
Check the Authorization bearer token and confirm the available balance in the console.
Context length exceeded
Prompt tokens exceed the DeepSeek V4 Pro context window. Trim or retrieve only the relevant evidence and reuse cached prefixes.
Rate limited (429)
Back off and retry with jitter; batch or queue requests instead of sending parallel bursts.
Content or tool call rejected
Review sensitive content, malformed tool-call arguments, and JSON schema mismatches before retrying.
How should you choose Chat Completions, Responses, and production traffic for DeepSeek V4 Pro?
The two protocols serve different workflows. This is a selection summary; use the existing API section and EvoLink documentation for exact request fields, and the existing Pricing section for token rates; count tool-call rounds into total task cost when budgeting.
01
Standard chat and client functions: Chat Completions
If you already use OpenAI-compatible chat, streaming, or client-side function calling, start with Chat Completions. Confirm message format, streaming completion, function arguments, and usage against what the current client expects.
Chat and client functions
02
Agent workflows: Responses
Evaluate the Responses API — added in the 0813 build — for longer multi-step agent flows and Codex-style integrations. Confirm supported request fields, failure states, and retry boundaries against the current EvoLink documentation before integration.
Agent workflows
03
Large-context work: watch cache and total cost together
Long documents, repositories, and conversations should not automatically put everything into one request. Compare representative context sizes, cache hits and misses, output length, and retries, then use the final charge to calculate successful-task cost.
Context cost
04
Production traffic: start small and keep a fallback
Send a small, observable set of tasks to DeepSeek V4 Pro first and retain a proven GPT, Claude, or Kimi route. EvoLink’s unified API centralizes model choice, usage, and balance, making it easier to change routes after rate limits, timeouts, schema errors, or tool failures.
Gradual rollout
Cache hits are billed at a fraction of fresh input
Prefix caching is automatic: a stable system prompt, repository instructions, and tool definitions are billed as cache hits on later requests instead of full input. Measure the real hit rate in your logs before estimating agent-loop cost.
1M-token context with a 384K output ceiling
Keep related files, specifications, and agent state in one working context, but treat the ceiling as capacity rather than a target: retrieve what matters, keep the prefix stable so it caches, and set an output budget per task.
Compare DeepSeek V4 Pro with Qwen 3.8 Max and GLM-5.3
EvoLink
Compare input/output rates, context, caching, and workload fit. Benchmark the same requests before choosing a route; prices below follow your account’s current rates.
Best forCoding agents, long task chains, tool-heavy workflows
What should you confirm before using DeepSeek V4 Pro in production?
Start with a small set of real workloads to determine whether the model meets your quality, latency, cost, and reliability needs before deciding which traffic should move.
Integration and usage data are clear
Confirm that the application uses the correct deepseek-v4-pro model ID and intended protocol, and that required responses, usage, and cache information are returned. Clear usage data supports cost analysis and gives Chat and Responses workflows a consistent observation method.
Outputs meet real business requirements
Test real code changes, long-document analysis, research, or structured extraction. Beyond answer quality, check whether tests pass, citations are reliable, function arguments are correct, and JSON Schema output can be consumed directly by downstream systems.
Latency and error handling meet expectations
Observe response time under representative traffic and prepare retry behavior for rate limits, timeouts, invalid structured output, and tool failures. Configurable model selection in EvoLink makes it easier to switch to a verified alternative when one route is temporarily unavailable.
Cost and model choice remain controllable
Use Pricing, usage, and final charges to calculate total cost for the same class of task, then decide whether DeepSeek V4 Pro belongs on the default route, difficult tasks, or fallback traffic. A unified gateway keeps quality and budget decisions separate from integration work.
Begin with a small, observable, reversible set of DeepSeek V4 Pro tasks. Expand only after quality, latency, and cost meet expectations. Keeping multiple model options behind the EvoLink unified API makes later scaling, switching, and cost optimization easier.
DeepSeek V4 Model Family
Same API key and balance — switch tiers without changing your integration.
DeepSeek V4.1 Flash
DeepSeek’s multimodal Flash model for text and image workloads
Is the DeepSeek V4 Pro API available through EvoLink now?
Yes. DeepSeek V4 Pro is live on the EvoLink production route and is not affected by DeepSeek’s September 14 change, which forwards deepseek-v4-pro to V4.1 Flash on DeepSeek’s own API. Send requests with the model ID deepseek-v4-pro on Chat Completions or Responses, and check the model ID, pricing, context, and supported workflows on this page before scaling up.
Is deepseek-v4-pro the API model ID?
Yes. The model ID you send in the request is deepseek-v4-pro — the same string as this page URL. Since August 13, 2026, that unchanged ID serves the upgraded 0813 GA build automatically; you do not change the ID to get the new build.
What changed in the DeepSeek V4 Pro 0813 build?
The 0813 GA build (released August 13, 2026) substantially improved agentic coding and terminal-task benchmarks over the April preview and added Responses API support. The callable model ID stays deepseek-v4-pro. The old aliases deepseek-chat and deepseek-reasoner were retired upstream on July 24, 2026.
What inputs and outputs does DeepSeek V4 Pro support?
Text input and text output only — DeepSeek V4 Pro has no vision capability on any protocol. On the Messages route, image and document content types are rejected; on Responses, image and file attachments are converted to placeholders rather than understood. Route screenshot or document-understanding tasks to a vision-capable model on the same EvoLink gateway.
Should I use Chat Completions or the Responses API?
Use Chat Completions for familiar chat, streaming, and client function flows. Evaluate Responses — added in the 0813 build — for longer-running agent workflows. The current API section and EvoLink docs remain the source for exact request fields and verified route support.
How should reasoning effort be selected?
Valid tiers are low, high, and max, and the default is high — medium is accepted but silently maps to high, so comparing medium against high produces no real difference. Start with low for routine work, compare low and high on the same tasks, and reserve max for the hardest problems where a failed attempt costs more than the extra reasoning tokens.
How do cached input and long agent loops affect DeepSeek V4 Pro pricing?
Cached input can reduce repeated-context cost, while cache misses, long outputs, retries, and repeated tool steps can make an agent loop more expensive than the headline input rate suggests. Use the live Pricing section and final request charges when calculating cost per accepted task.
What are the DeepSeek V4 Pro rate limits?
DeepSeek’s V4 documentation described an account-level concurrency cap on its direct API rather than per-token RPM or TPM limits — about 500 concurrent requests for Pro — with 429 returned beyond the cap and queued requests dropped after roughly 10 minutes. Those figures describe DeepSeek’s own API, not EvoLink route capacity. Measure the limits on your EvoLink account, keep in-flight requests below them, and use exponential backoff.
How should teams compare DeepSeek V4 Pro with GPT, Claude, or Kimi?
Run the same real tasks with consistent context, tools, and reasoning settings. Compare result quality, response time, token mix, tool performance, and total cost to decide which EvoLink route fits each traffic class.
Is DeepSeek V4 Pro open source?
The April Preview weights are MIT-licensed on Hugging Face; the 0813 GA build’s weights had not been published as of August 13, 2026. Open weights and the hosted API are independent access paths — the EvoLink route serves the hosted API, and the same MIT weights being available elsewhere is what makes third-party-host fallback possible.
Is there a DeepSeek V4 Pro Max model?
No. There is no separate “Pro Max” model — “max” is the top thinking-effort tier on the same deepseek-v4-pro model ID. If you see “Pro-Max” mentioned, it refers to running this model at maximum reasoning effort, not a different product.
What fallback should teams keep during rollout?
Keep a model that already handles the same workload reliably and leave route selection configurable. If rate limits, timeouts, or invalid output occur, EvoLink can route to GPT, Claude, Kimi, or another suitable alternative.
