DeepSeek: DeepSeek V4.1 Flash
deepseek-v4.1-flash- Input / Output
- $0.15 / $0.60
- Context
- 1M
- Max output
- 384K
- Modalities
- VisionChat
- Capabilities
- Tool usePrompt CacheReasoning
About DeepSeek V4.1 Flash
DeepSeek V4.1 Flash is the September 2026 update to DeepSeek's Flash model, and the first Flash release with native multimodal visual capability. It reads images together with text, reasons in thinking mode and calls tools, which makes it a fit for coding agents that need to look at an interface before acting. DeepSeek has retired earlier V4 Flash variants in its favor.
Where it works well
- Visual understanding is native, so a screenshot or diagram can be part of the same conversation as code and instructions.
- It combines vision with reasoning and tool calls, which lets an agent inspect a rendered page, interpret it and continue a multistep task.
- DeepSeek treats it as the current Flash model, so it is the version that receives the maker's updates and pricing changes for that line.
- The long context accepts large code and document sets next to images, and prompt caching covers repeated instructions.
When to choose another model
- Thinking is on by default and counts against max_tokens; a small budget can leave an empty body while the request is still billed.
- Schema-enforced JSON output is not available, so JSON-heavy pipelines should validate the output themselves.
- It is the small Flash tier; deepseek-v4-pro remains the DeepSeek choice for the hardest reasoning and coding work.
Getting started
Create API key
Create a key in Console, then use it with every model on the platform.
Send your first request
Copy the example for your language and run it against the endpoint.
Text to textResponses APIPOST/v1/responsesAPI format:curl https://api.tokenlab.sh/v1/responses \ -H "Content-Type: application/json" \ -H "Authorization: Bearer sk-xxx" \ -d '{ "model": "deepseek-v4.1-flash", "input": "Hello!" }'
Pricing
The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.
Rate window · Off-peak
per 1M tokens- Official price
- Input $0.15 / Output $0.60 / Cache read $0.003
- Verified price
- —
- Discount
- —
Rate window · Peak
per 1M tokens- Official price
- Input $0.30 / Output $1.20 / Cache read $0.006
- Verified price
- —
- Discount
- —
Cache read
- Official price
- $0.003
- Verified price
- —
- Discount
- —
| Official priceper 1M tokens | Verified priceper 1M tokens | Discount | |
|---|---|---|---|
| Rate window · Off-peak | Input $0.15 / Output $0.60 / Cache read $0.003 | — | — |
| Rate window · Peak | Input $0.30 / Output $1.20 / Cache read $0.006 | — | — |
| Prompt cache pricing | |||
| Cache read | $0.003 | — | — |
Usage & activity
Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.
Usage & availability
Last 24 hours- Requests
- Success rate
- P95 latency
- Total tokens
- 30-day success rate
- 99.7%
- 7-day median latency
- 8.3sn=24,545
- 7-day P95 latency
- 72.6sn=24,545
- 30-day requests
- 20K+
- Last active
- 3 minutes ago
Data is based on aggregate user requests, excluding status checks.
Open in Console
Open DeepSeek V4.1 Flash in Console with a prompt ready to edit or send.
Help me try deepseek-v4.1-flash with a short message at /v1/responses. Show the reply, latency, and cost.
Use cases
Best for- Vision
- Reasoning
UI-aware coding agents
Have the agent run the app, capture the screen, compare it with the design and edit the code until the layout matches.
Document pages as images
Read scanned forms, slides and chart-heavy reports directly, without a separate OCR step, and answer questions about them.
Bug reports with screenshots
Triage tickets that attach pictures: identify the screen, the visible error and the probable code path, then draft a fix.
Upgrade path for older Flash users
Move clients off retired V4 Flash variants and gain image input without changing request format.
Prompt examples
Open the attached screenshot of the settings page. Which controls overlap on narrow screens, and which stylesheet rule is responsible?
Read this scanned invoice image and return vendor, date, line items and total as JSON, marking any value that is unreadable.
Run the test command, view the failing snapshot image, and tell me whether the component or the snapshot is wrong before you edit anything.
This model has conditional pricing. Monthly token totals alone cannot produce a reliable estimate; use the detailed pricing for the request specification, cache, and applicable time window.
FAQ
What is new in DeepSeek V4.1 Flash?
Native multimodal visual capability. DeepSeek announced it on 10 September 2026 as the latest Flash model, with earlier V4 Flash variants retired and their legacy names temporarily redirected to it. It reads images while keeping reasoning and tool use.
Can DeepSeek V4.1 Flash see images?
Yes. Send images as content parts alongside text, and the model can describe, compare or act on what it sees. It handles screenshots, diagrams, scanned pages and charts in the same conversation as code and instructions.
Does it have a thinking mode?
Yes, and it defaults to thinking first. Thinking tokens count against max_tokens; if that budget is too small, the whole allotment can go to thinking and leave an empty body while the request is still billed. For long answers set max_tokens around 32000, or turn thinking off: add "thinking": {"type": "disabled"} (or "reasoning_effort": "none") to the request body, which works for both streaming and non-streaming requests. "enable_thinking": false and "reasoning": {"enabled": false} do not turn it off.
Should I use V4.1 Flash or V4 Pro?
Choose V4.1 Flash when the task includes images, when speed matters, or when you want the current Flash model. Choose V4 Pro for text-only work where reasoning depth decides success, such as large refactors or hard proofs.
Is DeepSeek V4.1 Flash the same as deepseek-v4-flash?
No. deepseek-v4-flash is the earlier text-only V4 Flash, while V4.1 Flash adds vision and is the line DeepSeek now names as its Flash model. Test your prompts again when you switch.
How much does DeepSeek V4.1 Flash cost?
On TokenLab, DeepSeek V4.1 Flash costs Input $0.15 / Output $0.60 per 1M tokens. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.
What are the context window and output limit of DeepSeek V4.1 Flash?
DeepSeek V4.1 Flash accepts up to 1,000,000 tokens of context and returns up to 384,000 tokens in one response.
Which endpoint should DeepSeek V4.1 Flash use?
Use https://api.tokenlab.sh/v1/responses for DeepSeek V4.1 Flash. The request example below shows the matching code shape.
Which operations does DeepSeek V4.1 Flash support?
DeepSeek V4.1 Flash supports Text to text. Select an operation above to see its endpoint and request example.
Compare DeepSeek V4.1 Flash
Guides that use DeepSeek V4.1 Flash
Sources
Reviewed Oct 15, 2026
