DeepSeek V4.1 Flash
Open-weights model released 10 September 2026. List price, context window and where to check them.
Price and size
USD list prices per million tokens. They leave out caching, batch discounts, tiered pricing and taxes.
Price and size
- Input
- $0.30per 1M tokens
- Output
- $1.20per 1M tokens
- Blended 3:1
- $0.523 parts input, 1 part output
- Context window
- 1Mtokens
- Typical month
- $6.0010,000 requests of 1,000 in, 250 out
The record
The record
- Provider
- DeepSeek
- Weights
- Open weights. Read the licence before commercial use: we have not reviewed it.
- Inputs it accepts
- text, image
- Release date
- 10 September 2026
- Maximum output
- 384K tokens
What we do not know
Stating the gaps matters as much as stating the figures.
Unknowns
- Quality benchmarks (none published here)
- Licence terms
- Speed and reliability
Where to check these figures
These are the provider's own pages. Listing a link here does not mean we have checked it.
Sources
Data for every model: /data/. How we handle model figures: methodology.
Frequently asked questions
How much does DeepSeek V4.1 Flash cost?
The supplied list price is $0.30 per million input tokens and $1.20 per million output tokens. At 10,000 requests a month of 1,000 input and 250 output tokens, that is about $6.00 a month. It leaves out caching, batch discounts, tiered pricing and taxes. Check DeepSeek's pricing page before you budget.
How large is the DeepSeek V4.1 Flash context window?
The supplied context window is 1M tokens. That figure has not been re-verified against DeepSeek's documentation.
Is DeepSeek V4.1 Flash an open-weights model?
Yes, it is recorded as open weights, so you may be able to self-host it. Read the licence before commercial use: we have not reviewed it.
Is DeepSeek V4.1 Flash better than other models?
We cannot say. No quality benchmark for it is published on this site, and the Model Finder orders results by cost, not quality. Use the finder to shortlist models, then test them on your own prompts.