DeepSeek V4 Flash Vision Exp is an experimental vision-enabled build of DeepSeek V4 Flash from DeepSeek. It adds image understanding on top of the base model while matching it on text tasks, carrying forward the same 1M-token context window and thinking mode enabled by default.
| Developer | DeepSeek |
|---|---|
| Model ID | deepseek/deepseek-v4-flash-vision-exp |
| Release Date | August 21, 2026 |
| Context Window | 1,048,576 tokens (~1M) |
| Max Output Tokens | 384,000 |
| Modalities | Image, Text (input) → Text (output) |
| Capabilities | Reasoning, tool/function calling, streaming, structured (JSON schema) output, vision |
Image input is layered on top of DeepSeek V4 Flash without giving up its text performance, so the same model works for plain chat and for tasks that require reading an image.
Reasoning tokens are generated automatically before the final answer, which helps on multi-step or ambiguous requests without extra configuration.
Long documents, transcripts, or multi-turn conversations can be passed in a single request without chunking.
Ask questions about screenshots, diagrams, or photos alongside regular text instructions in the same request.
The 1M-token window and default thinking mode suit multi-step analysis over long inputs.
As an experimental build, it's a way to evaluate DeepSeek's upcoming vision capabilities before a general release.
Measured on the AI/ML API gateway across live traffic. Figures are medians unless stated. Last updated 2026-10-12.
This model is documented for visual understanding workflows, including describing pictures, reading text from screenshots, and analyzing charts.
It is also documented for multimodal agent workflows, where visual understanding is combined with tools to support practical agent-based tasks.
It is documented for UI analysis and screenshot-based tasks, such as identifying interface elements and reasoning from visual evidence.
It is documented for document extraction from images, including invoices, receipts, delivery notes, and insurance forms.
import requests r = requests.post( "https://api.aimlapi.com/v1/chat/completions", headers={"Authorization": "Bearer " + AIMLAPI_KEY}, json={ "model": "deepseek/deepseek-v4-flash-vision-exp", "messages": [ { "role": "user", "content": "Hello!" } ] }, ) print(r.json())
const r = await fetch("https://api.aimlapi.com/v1/chat/completions", { method: "POST", headers: { Authorization: `Bearer ${process.env.AIMLAPI_KEY}`, "Content-Type": "application/json", }, body: JSON.stringify({ "model": "deepseek/deepseek-v4-flash-vision-exp", "messages": [ { "role": "user", "content": "Hello!" } ] }), }); console.log(await r.json());
curl -X POST https://api.aimlapi.com/v1/chat/completions \ -H "Authorization: Bearer $AIMLAPI_KEY" \ -H "Content-Type: application/json" \ -d '{"model":"deepseek/deepseek-v4-flash-vision-exp","messages":[{"role":"user","content":"Hello!"}]}'
OpenAI-compatible — swap the base URL and it works with your existing SDK.
| Type | Price |
|---|---|
| Input | |
| Output | |
| Cached input |
Reasoning tokens are billed at the same rate as output tokens.
| Model | Input | Output | Context | Best for |
|---|---|---|---|---|
DeepSeek V4 Flash Vision Exp This page | Vision + long-context reasoning | |||
| Reasoning + agents | ||||
| Balanced coding + agents | ||||
| Long-context, multimodal & agentic workflows | ||||
| Reasoning + agents |
DeepSeek V4 Flash Vision Exp has a 1,048,576 tokens context window and can return up to 384,000 tokens.
DeepSeek V4 Flash Vision Exp takes image, text as input and returns text.
Use deepseek/deepseek-v4-flash-vision-exp as the model id. Requests go to https://api.aimlapi.com/v1/chat/completions.
DeepSeek V4 Flash Vision Exp became available on August 21, 2026.
DeepSeek V4 Flash Vision Exp is priced at input $0.572 / 1M tokens, output $1.716 / 1M tokens, cached input $0.0182 / 1M tokens.
Yes, DeepSeek V4 Flash Vision Exp can stream responses as they are generated.
Yes, DeepSeek V4 Flash Vision Exp accepts image input alongside text.
DeepSeek V4 Flash Vision Exp was built by DeepSeek.
DeepSeek V4 Flash Vision Exp is described as an experimental build, adding vision capability on top of DeepSeek V4 Flash.
Yes, it matches DeepSeek V4 Flash on text tasks while adding image understanding.
It includes reasoning as a core feature, with thinking mode enabled by default.
