DeepSeek V4 Flash pricing
DeepSeek V4 Flash is an API model from DeepSeek. It costs $0.30 per million input tokens and $1.20 per million output tokens, with a 1M-token context window. At a blended $0.52 per million tokens it ranks #8 cheapest of 46 tracked models (budget band).
- Input / 1M tokens
- $0.30
- Output / 1M tokens
- $1.20
- Cached input / 1M
- $0.006
- Blended (3:1)
- $0.52
- Context window
- 1M
- Max output
- 393K
- Vision · Tools · Reasoning
- Yes · Yes · Yes
- Listed since
- Jun 20, 2026
What it costs
| Workload | Per 1,000 requests |
|---|---|
| Chatbot reply2,000 input + 500 output tokens | $1.20 |
| RAG / search answer8,000 input + 700 output tokens | $3.24 |
| Long-document summary100,000 input + 2,000 output tokens | $32.40 |
Price history
| First observed | Input | Output | Change |
|---|---|---|---|
| Sep 13, 2026 | $0.30 | $1.20 | -20% blended |
| Aug 23, 2026 | $0.44 | $1.32 | +277% blended |
| Jun 20, 2026 | $0.14 | $0.28 | listed |
Dates are the week a price first appeared in the public price list we track, not necessarily the provider's announcement date.
Compare
Similarly priced
- MiniMax M3$0.52 blended
- Codestral (2508)$0.45 blended
- GPT-5 mini$0.69 blended
- Mistral Large 3$0.75 blended
DeepSeek V4 Flash in the news
Prices are the provider's list prices for its own API (USD), taken daily from LiteLLM's public model price list. Discounts for batch processing, caching or committed use, and regional or cloud-marketplace pricing, aren't reflected. Check the provider before estimating production costs. See also the TW Model Price Index.