DocsSecuritySign in
    Get started
    Deepseek logoDeepseek

    deepseek-flash

    Auto-router

    deepseek/deepseek-flash

    ChatVisionFilesToolsSystem promptReasoning
    Use this model

    Context

    1M

    Max output

    384K

    Input / 1M

    $0.30

    Output / 1M

    $1.20

    Cached / 1M

    $0.006

    Cutoff

    unknown

    About this model

    DeepSeek-V4.1-Flash (deepseek-flash) is DeepSeek fast general-purpose model with 1M token context and 384K max output, supporting vision, tool calls, JSON mode, and thinking mode. Standard (peak-hour) pricing shown; off-peak discount (50% off, 01:00-04:00 and 06:00-10:00 UTC weekdays) also available. Supersedes the retired deepseek-v4-flash and deepseek-v4-flash-vision-exp, which are now routed to this model.

    Best suited for

    • Multimodal agent tasks requiring visual understanding
    • Long-context conversations (up to 1M tokens)
    • Coding and agentic workflows
    • Screenshot/image-grounded reasoning

    Capabilities

    Vision

    Accepts images alongside text in the same message.

    Files

    Accepts file attachments — PDFs, transcripts, spreadsheets.

    Tools

    Native function calling, so agents can invoke your endpoints.

    System prompt

    Honours a dedicated system role, separate from the user turn.

    Reasoning

    Emits a separate thinking pass before the answer.

    Supported parameters

    max_tokensMax Tokens Limit

    Specifies the maximum number of text units (tokens) allowed in a response, limiting its length.

    parallel_tool_calls

    Enables parallel execution of tools, allowing multiple tools to run simultaneously.

    reasoning

    Controls the thinking effort level used by the model (non-thinking or thinking mode with low/high/max effort).

    response_type

    Defines the format or type of the generated response.

    stream

    Sends the response in real-time as it's being generated.

    tool_choice

    Decides whether to use tools or just the model for generating responses.

    tools

    Lists tool definitions or capabilities available to the model.

    Use this model

    OpenAI-SDK compatible, with gateway fallback and routing across providers.

    Other Deepseek models

    ModelContextMax outputInputOutputCapabilities
    deepseek-v4-pro

    Chat

    		1M384K$1.32$3.96Tools, System prompt, Reasoning

    Pricing & provider details

    Deepseek logo

    Deepseek

    deepseek/deepseek-flash

    Input · per 1M
    $0.30
    Output · per 1M
    $1.20
    Cached input · per 1M
    $0.006
    Context window
    1,000,000 tokens
    Max output
    384,000 tokens
    Knowledge cutoff
    unknown
    Auto-router
    Supported

    Start building today

    Route deepseek-flash — and every other model in the catalogue — through one endpoint, with failover built in.

    Get started freeAll models