Price Index · All models · DeepSeek
Official list price · checked 20 November 2026
DeepSeek Flash API pricing (November 2026)
DeepSeek charges $0.30 per million input tokens and $1.20 per million output tokens for DeepSeek Flash, according to its official pricing page. This price has applied since 12 November 2026 in our records.
What DeepSeek Flash costs for common jobs
Per 1,000 requests, at list price. Batch and cached-input discounts would lower these.
| Job | Tokens per request | Cost per 1,000 requests |
|---|---|---|
| Chat reply | 1,000 tokens in, 300 out | $0.66 |
| Question over documents | 5,000 in, 500 out | $2.10 |
| Long-report summary | 50,000 in, 1,000 out | $16.20 |
Price history
| Price read on | Input | Output | Source |
|---|---|---|---|
| 12 November 2026 | $0.30 | $1.20 | Pricing page |
The page is re-read every day. History is append-only: a price change adds a new dated row and never overwrites an old one.
Compared with similarly priced models
| Model | Input | Output | Blended | vs DeepSeek Flash |
|---|---|---|---|---|
| Amazon Nova 2 LiteAmazon | $0.30 | $2.50 | $0.85 | +62% |
| Qwen3.7 PlusAlibaba (Qwen) | $0.40 | $1.60 | $0.70 | +33% |
| DeepSeek FlashDeepSeek | $0.30 | $1.20 | $0.525 | — |
| Command RCohere | $0.15 | $0.60 | $0.2625 | -50% |
| Mistral Small 4Mistral | $0.15 | $0.60 | $0.2625 | -50% |
| GLM-5.3-FlashZ.ai (Zhipu) | $0.15 | $0.50 | $0.2375 | -55% |
| Qwen3.8 FlashAlibaba (Qwen) | $0.15 | $0.47 | $0.23 | -56% |
Other DeepSeek models
- DeepSeek V4 Pro $1.32 / $3.96
Questions
How much does DeepSeek Flash cost?
DeepSeek lists DeepSeek Flash at $0.30 per million input tokens and $1.20 per million output tokens on its official pricing page (standard pay-as-you-go tier, uncached input, short-context rate), as read on 20 November 2026.
Has DeepSeek Flash's price changed?
Not since the index first recorded it on 12 November 2026. The price is re-read from DeepSeek's page every day, and any change is added as a new dated record.
What does DeepSeek Flash cost for 1,000 typical requests?
About $0.66 for 1,000 chat replies of 1,000 input and 300 output tokens each, at list price before any batch or caching discount.
Is there a cheaper model at a similar level?
Among similarly priced models from other providers, Command R (Cohere) lists at $0.2625 blended per million tokens, against $0.525 for DeepSeek Flash. Price is not quality: compare them on your own task.
Put this price on your site
Paste this into a page, doc or README. It shows DeepSeek Flash's current list price and updates when the price changes. Free to use.
<iframe src="https://aicapitalindex.com/embed/deepseek-flash" title="DeepSeek Flash API price" width="360" height="130" style="border:0" loading="lazy"></iframe>
How this price is read
The list price for standard, pay-as-you-go use: uncached input, the short-context tier, and the peak rate where prices vary by time of day. Batch, cached and enterprise discounts are excluded. Blended price weights input and output 3 to 1. Notes recorded with this price: PEAK rate, cache-miss input. Off-peak $0.15/$0.60.
Data: Certain Model Price Index, CC BY 4.0. Check the provider's page before relying on a price.