HTTP 200 verified twice
Finding your next stop…
Loading the latest directory information.
Loading the latest directory information.
DeepSeek API offering deepseek-v4-flash, deepseek-v4-pro, and deepseek-v4-flash-vision-exp models with OpenAI- and Anthropic-compatible endpoints, function/tool calling, JSON output, vision input, thinking mode, and FIM code completion.
Start here for the decision-making essentials: what DeepSeek V3 does, who it is for, how it is accessed, and the first-party sources behind this profile.
Files API is free to use
images billed at V4-Flash pricing per token
Read-only public captures of DeepSeek V3’s homepage. Screenshots are dated, never live embeds, and open full-screen.
Published specifications and use cases for DeepSeek V3. Repository dates are kept separate from the model's original release date.
These are publisher-reported results. Test conditions and benchmark versions can differ.
Short answers to the questions buyers and builders commonly ask about DeepSeek V3. Each answer cites the shared ledger below, where every source is listed once.
OpenAI- and Anthropic-compatible API; works with OpenAI/Anthropic SDKs and compatible soft · Context Caching on Disk technology enabled by default for all users, with prompt_cache_hit · Files API allows uploading images once and referencing them later by file_id across multip · External tool calling via the DeepSeek API
The DeepSeek API uses an API format compatible with OpenAI/Anthropic. By modifying the configuration, you can use the OpenAI/Anthropic SDK or softwares compatible with the OpenAI/Anthropic API to access the DeepSeek API.
Code completion · Content completion
FIM is commonly used for content completion、code completion.
The max tokens of FIM completion is 4K · For multi-turn conversations, the user must concatenate all previous conversation history
The max tokens of FIM completion is 4K.
Files API is free to use; images billed at V4-Flash pricing per token
Files API is now live 📁 🔹 Free to use 🔹 Upload an image once, then reference it by file_id to save request bandwidth
OpenAI-compatible endpoint at https://api.deepseek.com supporting Chat Completions, Messag · OpenAI-compatible Chat Completions endpoint at base_url https://api.deepseek.com · Anthropic-compatible Messages endpoint at base_url https://api.deepseek.com/anthropic · OpenAI Responses API format supported at base_url https://api.deepseek.com, enabling integ
The same file operations are also available through the Anthropic-compatible endpoint, with base_url = https://api.deepseek.com/anthropic.
DeepSeek Harness 0.1.1 with out-of-the-box support for deepseek-v4-flash-vision-exp · Native integration with Claude Code via Anthropic-compatible API · Continue (VSCode plugin) integration supports code completion via the API
DeepSeek Harness 0.1.1 was released today with out-of-the-box support for the new model.
This profile connects the jobs DeepSeek V3 is described as handling with its delivery model, access options and the subjects used to match it to related products in this directory.
Each fact points to a recorded source, making it easy to distinguish verified product information from claims that need checking.
HTTP 200 verified twice
Your First API Call | DeepSeek API Docs
OpenAI- and Anthropic-compatible API at https://api.deepseek.com and https://api.deepseek. · deepseek-v4-flash model (version DeepSeek-V4-Flash-0731) · deepseek-v4-pro model (version DeepSeek-V4-Pro-0813)
OpenAI- and Anthropic-compatible API; works with OpenAI/Anthropic SDKs and compatible software by changing the base URL
Function/tool calls supported on all listed models; up to 128 functions with optional strict-mode schema enforcement
JSON Output supported on all listed models via response_format type=json_object
Vision input (image + text) on deepseek-v4-flash-vision-exp; supported file types JPEG, PNG, GIF, WebP
deepseek-v4-flash, version DeepSeek-V4-Flash-0731
deepseek-v4-pro, version DeepSeek-V4-Pro-0813
deepseek-v4-flash-vision-exp (experimental multimodal model)
2026/08/21
Multimodal: mixed text + image input; images provided via base64, external URLs, or the Files API
A concise view of the jobs, capabilities and integrations described in the recorded product sources.
Documented product formats, platforms and official distribution destinations. Availability can vary by region and plan.
A concise history of software releases and material product changes. Events appear only when a dated source supports what changed.
Showing the newest updates and meaningful milestones. Open an entry for its summary and source.
Preview of DeepSeek-V4-Pro (1.6T total/49B active params) and DeepSeek-V4-Flash (284B total/13B active params), both open-sourced with API access. Introduces token-wise compression plus DSA, making 1M context the default across official services.
View source [19]Official successor to V3.2-Exp, available on App, Web and API alongside V3.2-Speciale (API-only until Dec 15, 2025). First DeepSeek model to integrate thinking directly into tool-use, with tool-use supported in both thinking and non-thinking modes.
View source [18]Experimental model built on V3.1-Terminus introducing DeepSeek Sparse Attention (DSA) for faster, more efficient long-context training and inference. API prices cut by 50%+ effective immediately, with V3.1-Terminus remaining available until Oct 15, 2025.
View source [17]Facts, answers, structured details, milestones and primary resource links cite this shared ledger. Each external page appears once; release tags from the same GitHub project are grouped under one release history.
These tools share a workflow, capability or audience with DeepSeek V3. They may complement it rather than replace it, and are not presented as integrations or endorsements.
JSON Output by setting response_format to {'type': 'json_object'} and including the word 'json' in the prompt
Context Caching on Disk technology enabled by default for all users, with prompt_cache_hit_tokens and prompt_cache_miss_tokens reported per request
Files API allows uploading images once and referencing them later by file_id across multiple requests to the deepseek-v4-flash-vision-exp model
OpenAI-compatible endpoint at https://api.deepseek.com supporting Chat Completions, Messages, and Responses; plus Anthropic-compatible endpoint at https://api.deepseek.com/anthropic
Works smoothly across agent frameworks, combining visual understanding with a wide range of tools
DeepSeek Harness 0.1.1 with out-of-the-box support for deepseek-v4-flash-vision-exp
Reported multimodal agent performance close to Opus-4.8 (vendor-reported comparison vs. V4-Flash)
Files API is free to use; images billed at V4-Flash pricing per token
External tool calling via the DeepSeek API
Thinking mode outputs a chain-of-thought before the final answer
Thinking mode supports tool calls starting from DeepSeek-V3.2
Strict JSON Schema mode for function tool calls (Beta), with supported types object/string/number/integer/boolean/array/enum/anyOf
OpenAI-compatible Chat Completions endpoint at base_url https://api.deepseek.com
Anthropic-compatible Messages endpoint at base_url https://api.deepseek.com/anthropic
OpenAI Responses API format supported at base_url https://api.deepseek.com, enabling integration with Codex
deepseek-v4-pro
deepseek-v4-flash
deepseek-v4-flash-vision-exp (vision-capable model)
Image input supported via deepseek-v4-flash-vision-exp (JPEG/PNG/GIF/WebP, http(s) URL or base64 data URL or Files API file_id)
Native integration with Claude Code via Anthropic-compatible API
Fill-in-the-Middle (FIM) completion (Beta) where users provide a prefix and optional suffix and the model completes the content in between
Stateless /chat/completions API for multi-turn conversations
Chat Prefix Completion (Beta) where users provide an assistant's prefix message for the model to complete
Code completion
Content completion
The max tokens of FIM completion is 4K
Beta features are accessed via base_url=https://api.deepseek.com/beta
Standard endpoint base_url is https://api.deepseek.com
