DeepSeek V4 Flash legacy API name
DeprecatedA still-accepted legacy API name whose requests are now served by DeepSeek-V4.1-Flash; the corresponding V4 Flash model is retired.
Access details: DeepSeek API
An alias. The provider can point it at a newer version.
deepseek-v4-flashUpdated Sep 13, 2026
Capabilities & limits
- Context window
- 1M
- Maximum output
- 384K
- Image input
- Supported
- Tool calling
- Supported
- Structured output
- Supported
Pricing
USD / 1M tokens- Input
- $0.30
- Output
- $1.20
Standard rates per million tokens. Discounts and long-context rates, where offered, are on DeepSeek's pricing page.
Full rate card 6 rates
All prices in USD.
Cache-hit input, off-peak
$0.003 per 1M tokens- Off-peak: all times outside 01:00-04:00 and 06:00-10:00 UTC, Monday-Friday
Cache-hit input, peak
$0.006 per 1M tokens- Peak: 01:00-04:00 and 06:00-10:00 UTC, Monday-Friday
Cache-miss input, off-peak
$0.15 per 1M tokens- Off-peak: all times outside 01:00-04:00 and 06:00-10:00 UTC, Monday-Friday
Cache-miss input, peak
$0.30 per 1M tokens- Peak: 01:00-04:00 and 06:00-10:00 UTC, Monday-Friday
Output, off-peak
$0.60 per 1M tokens- Off-peak: all times outside 01:00-04:00 and 06:00-10:00 UTC, Monday-Friday
Output, peak
$1.20 per 1M tokens- Peak: 01:00-04:00 and 06:00-10:00 UTC, Monday-Friday
Legacy name deepseek-v4-flash: requests are served by DeepSeek-V4.1-Flash and billed at the Flash (deepseek-flash) price. Off-peak rates are half the peak rates; peak hours exclude Chinese public holidays, which are off-peak in full. These rates took effect at 04:00 UTC on 2026-09-10.
Price history
Sep 10, 2026
Input $0.30 per million tokens
Sep 10, 2026
Output $1.20 per million tokens
Technical details
- Parameters
- 552B
- Active parameters
- 8B per token in prefill, 16B in decode
- Architecture
- Mixture-of-Experts with a Causal Encoder-Decoder (CED) architecture
- Training / knowledge cutoff
- Not published
- Input modalities
- Text, image
- Output modalities
- Text
From DeepSeek
Official announcement ↗How they describe it
Your fast, efficient, and economical choice.
What they say is new
Reasoning capabilities closely approach V4-Pro.
Release & lifecycle
- First available
- Apr 24, 2026
- available
- As of Sep 13, 2026
- deprecated
- As of Oct 1, 2026