DeepSeek's flagship V4 model - a preview from the Chinese AI lab. Hybrid thinking / non-thinking mode, 1M context window, 384K max output, frontier-class agentic capabilities.
Context window
1M
tokens
Parameters
not disclosed (successor to V3/V3.1 MoE with 671B total / 37B activated)
parameters
Max output
384,000
tokens
Release date
1 July 2026
Access:APIHostedDeployment:☁ Cloud
Overview
Access & deployment
APIHosted
Cloud
Weights: Closed
Key parameters
📏 Context: 1M
🧩 Parameters: not disclosed (successor to V3/V3.1 MoE with 671B total / 37B activated)
✓ Tools
📥 Input: text, structured data, documents
Technical specification
Context window
1M
tokens
Parameters
not disclosed (successor to V3/V3.1 MoE with 671B total / 37B activated)
parameters
Max output tokens
384,000
tokens per response
Knowledge cutoff
1 Mar 2026
Knowledge boundary
License
Proprietary (DeepSeek Open Platform commercial). Ważne: wagi V4 nie zostały opublikowane w przeciwieństwie do wcześniejszych modeli V3/R1 (MIT/DeepSeek License).
Hardware requirements
The model is available exclusively via the DeepSeek Open Platform (platform.deepseek.com) and chat.deepseek.com - V4 weights are not released for self-hosting (as of July 2026). Earlier DeepSeek models (V3, R1) are available open-weights on GitHub.
Features:✓ Tool use
Modalities
⬇ Input
textstructured_datadocuments
⬆ Output
textcodestructured_data
Capabilities and applications
Native model capabilities
Advanced reasoning
The ability to perform multi-step, structured reasoning: analysing problems, planning steps, and drawing conclusions from hypotheses. Reasoning-first models (e.g. GPT-5.1 Thinking) dedicate a portion of inference to chains of thought before responding.
Category: reasoning
Extended thinking mode
A reasoning-model variant with a larger inference budget: more thinking cycles, higher answer precision at the cost of response time. Choice between 'standard' and 'extended' thinking is left to the user (e.g. the selector in GPT-5.2 Pro).
Category: reasoning
Adaptive reasoning effort
The model decides how much 'thinking' to allocate to a given query: simple questions are answered quickly, complex problems receive more inference cycles. A GPT-5.1 feature (both Instant and Thinking) that shortens time on easy tasks and extends it for hard ones.
Category: reasoning
Multi-step reasoning
Carrying out multi-step chains of reasoning across long, complex tasks.
Category: reasoning
Mathematical reasoning
The model's ability to solve mathematical tasks requiring multi-step reasoning — equations, proofs, combinatorics, geometry, calculus and competition-level problems.
Category: reasoning
Coding
Generating, analysing and modifying code in many programming languages. Covers writing functions, debugging, refactoring, code review, and creating tests. Measured by benchmarks such as HumanEval and SWE-bench.
Category: coding
Agentic coding
Multi-hour, multi-step programming tasks performed autonomously by the model: cloning a repository, running tests, iterating on fixes, integrating with CLI tools. Characteristic of Codex variants (GPT-5.1-Codex-Mini, Codex-Max).
Category: coding
Agentic capability
The model's ability to autonomously plan and execute multi-step tasks by sequentially using tools, maintaining context, and adapting to intermediate results.
Category: planning
Tool use
The model's ability to call external functions, APIs and tools during a conversation: calculator, search engine, code editor, database. The model decides when and how to use a tool and interprets its result.
Category: planning
Long context
Support for large context windows — tens to hundreds of thousands (or millions) of input tokens. Enables analysis of entire codebases, long documents, and many parallel conversations without losing earlier information. GPT-5.1 supports 400,000 tokens.
Category: language
Prompt caching
Cost-performance optimisation: repeated prompt fragments (e.g. system prompt, long documentation) are cached server-side and cheaper in subsequent calls. Significantly reduces cost for applications with long contexts.
Category: other
Multilingual
Competence in many natural languages (from a few to over a hundred): understanding, generation, translation, and code-switching within a single conversation. Frontier models support a wide range of languages with comparable quality.
Category: language
Pricing
Technical architecture
Core Architecture
Deployment and security
🔒 Security / Enterprise
✓ Verified enterprise information
The model is hosted on DeepSeek's infrastructure in China (Hangzhou). Due to the company's PRC headquarters, the model is subject to Chinese AI regulations (including politically sensitive content filters required by the Cyberspace Administration of China). V4 weights are not available for offline deployment - companies with data residency requirements outside China should account for this. DeepSeek maintains a privacy policy compliant with China's PIPL.