DeepSeek V4.1 Flash API

DeepSeek V4.1 Flash: very low-cost fast model with a 1M-token context.

DeepSeek·ChatAvailable

from$0.018/ 1M input tokens

Model IDdeepseek-v4.1-flash

fasttools1m-context

Input

Text

Output

Chat

Provider

8 providers

Last 7 days · not enough recent requests for stats yet.

DeepSeek V4.1 Flash pricing

Prices are per unit; requests without a provider are routed to the cheapest provider that supports your parameters.

OptionPrice (USD)
All options
$0.9 / 1M output tokens
90 Credits

Try DeepSeek V4.1 Flash before you integrate

Test output quality, compare provider prices and estimate cost before writing any code.

Provider
Chat

Settings

Price01router
1M input tokens$0.018
1M output tokens$0.9

Credits for max_tokens are held first, then settled to the actual usage.

DeepSeek V4.1 FlashAuto (cheapest)
DeepSeek V4.1 Flash

Start a conversation. Replies stream in and are billed by actual token usage.

Enter to send · Shift + Enter for a new line

DeepSeek V4.1 Flash prompt examples

Ready-to-use prompts. Load one into the playground and tweak it.

Source
Scene

19 prompts

List every risky change in this diff with file and line, and explain what could break. <paste diff here>

Read this scanned invoice and return vendor, date, line items and total as JSON.

Bulk ticket tagging

Classification

Tag each of these support tickets with a product area and an urgency from 1 to 3. Return CSV. <paste tickets here>

DeepSeek V4.1 Flash API reference

Endpoints, request fields and runnable examples for this model.

Base URL

https://01router.com/api/v1
llms.txt

Quick start

Set API_KEY to your project key, then run a sample from your server. Replace example media URLs with your own public URLs.

Auto-routing chooses a provider that supports the request. Select a provider to see its input constraints and pinned request examples.

curl https://01router.com/api/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4.1-flash",
    "messages": [{ "role": "user", "content": "Hello!" }],
    "stream": false
  }'

Endpoints

Only endpoints relevant to this model are listed. Paths below include /api/v1. Click an endpoint to read its usage.

Authentication

Send your API key as a Bearer token. Requests are billed to the key owner's credits.

Authorization: Bearer YOUR_API_KEY
Content-Type: application/json
Get an API key

Input

Send messages directly in the request body. Chat requests do not use an input wrapper.

Request body

NameTypeRequiredDescription
modelstringYes"deepseek-v4.1-flash"
providerstringNoPin a provider; omit to auto-route to the cheapest. One of: 01router | wavespeed | kie | apimart | evolink | atlascloud | vercel | openrouter
messagesobject[]YesOpenAI-format messages, each with role and content.
streambooleanNotrue returns an SSE stream.
max_tokensintegerNoOutput token cap; credits are pre-held for it (default 4096).

Output

This example uses stream: false. Read the reply from choices[0].message.content. With stream: true, read SSE chunks from choices[].delta until [DONE].

Task completed

{
  "id": "chatcmpl-example",
  "object": "chat.completion",
  "model": "deepseek-v4.1-flash",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hello! How can I help?"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 9,
    "completion_tokens": 8,
    "total_tokens": 17
  }
}

Response fields

NameTypeRequiredDescription
idstringYesChat completion ID
objectstringYeschat.completion
modelstringYesdeepseek-v4.1-flash
choicesobject[]YesReply choices; message.content contains the non-streaming response.
usageobjectNoToken usage: prompt_tokens, completion_tokens and total_tokens. Availability depends on the provider.

Model schema

This public endpoint returns the model parameters and each provider’s supported constraints and prices.

curl "https://01router.com/api/v1/models/deepseek-v4.1-flash"

Errors

Check the HTTP status and error.message. Fix the request before retrying validation or authentication errors. A failed generation is reported in status and error on the task itself.

{
  "error": {
    "message": "Invalid or missing API key",
    "code": "unauthorized"
  }
}

400Invalid body or parameter value

401Missing or invalid API key

402Not enough credits

404Unknown model or generation

429Too many requests — retry after the given seconds

502Every provider rejected the task (credits refunded)

Auth, routing, billing and error codes are covered in →

Use with AI

Give your coding assistant this model reference: endpoints, input constraints, output fields and request examples in plain text.

Open llms.txt

How does DeepSeek V4.1 Flash compare?

Specs and starting prices side by side. Test with the same prompt before you choose.

DeepSeek V4.1 FlashGPT-6 AstraGPT-6 SolGPT-6 Luna
VendorDeepSeekOpenAIOpenAIOpenAI
Starting price$0.018 / 1M input tokens$2.52 / 1M input tokens$0.54 / 1M input tokens$0.027 / 1M input tokens
InputTextTextTextText
Provider8 providers7 providers7 providers8 providers

DeepSeek V4.1 Flash API FAQ

What changed from DeepSeek V4 Flash?

V4.1 Flash adds native image understanding (V4 Flash was text-only) and a more efficient mixture-of-experts architecture with a much smaller memory footprint.

What are the limits?

About 1M tokens of context and up to about 384K output tokens. It accepts text and images (PNG, JPEG, WebP, GIF).

Is thinking on by default?

On DeepSeek's own API thinking is on by default, with a non-thinking mode for faster answers. Available controls vary by provider.

What can I build with it?

Coding assistants, terminal agents, OCR and screenshot reading, chart extraction, and high-volume classification or extraction where cost matters.

How is it billed on 01router?

Per 1M input and output tokens. Provider prices differ a lot for this model, so compare them in the pricing table; requests without a provider go to the cheapest one.

Related models

Other models you can call with the same API key.