DeepSeek V4 Flash
DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window.
deepseek-v4-flashReleased 24 Apr 2026 (Apr 2026)Pricing
List prices at DeepSeek, in US dollars per million tokens.
| Tier | Input / 1M | Output / 1M | |
|---|---|---|---|
| Standard | $0.30 | $1.20 | |
| Cached input 98% off input | $0.006 | 98% off input | |
| Cache write, 5 minutes | $0.00 |
What it costs a month
At DeepSeek list prices, 30 days.
- Chat assistant5,000 requests a day, 2K in / 0.5K out, cached$154
- RAG over documents2,000 requests a day, 8K in / 0.6K out, cached$117
- Coding agent1,000 requests a day, 40K in / 2K out, cached$185
- Overnight batch20,000 requests a day, 4K in / 0.4K out$1,008
Where to run it
30 hosts serve DeepSeek V4 Flash. The comparison is on blended price, 3 parts input to 1 part output.
| Host | In / out per 1M | Model id | |
|---|---|---|---|
DeepSeekLab API deepseek-v4-flash | $0.30 / $1.20 | deepseek-v4-flash | |
| $0.09 / $0.18 −79% vs headline | DeepSeek-V4-Flash-0731 | ||
DeepInfraInference DeepSeek-V4-Flash | $0.09 / $0.18 −79% vs headline | DeepSeek-V4-Flash | |
GMI CloudInference DeepSeek-V4-Flash | $0.091 / $0.182 −78% vs headline | DeepSeek-V4-Flash | |
PinstripesCloud deepseek-v4-flash | $0.10 / $0.20 −76% vs headline | deepseek-v4-flash | |
Kyma APICloud deepseek-v4-flash | $0.123 / $0.246 −71% vs headline | deepseek-v4-flash | |
| $0.13 / $0.26 −69% vs headline | DeepSeek-V4-Flash-0731 | ||
| $0.13 / $0.26 −69% vs headline | deepseek-v4-flash-0731 | ||
ai&Cloud deepseek-v4-flash | $0.15 / $0.25 −67% vs headline | deepseek-v4-flash | |
TensormeshCloud DeepSeek-V4-Flash | $0.14 / $0.28 −67% vs headline | DeepSeek-V4-Flash | |
| $0.14 / $0.28 −67% vs headline | databricks-deepseek-v4-flash-0731 | ||
StreamLakeCloud deepseek-v4-flash | $0.14 / $0.28 −67% vs headline | deepseek-v4-flash | |
Tencent TokenHubCloud deepseek-v4-flash | $0.14 / $0.28 −67% vs headline | deepseek-v4-flash | |
Nebius AI StudioInference DeepSeek-V4-Flash | $0.14 / $0.28 −67% vs headline | DeepSeek-V4-Flash | |
W&B InferenceInference DeepSeek-V4-Flash | $0.14 / $0.28 −67% vs headline | DeepSeek-V4-Flash | |
| $0.14 / $0.28 −67% vs headline | DeepSeek-V4-Flash-0731 | ||
Fireworks AIInference deepseek-v4-flash | $0.14 / $0.28 −67% vs headline | deepseek-v4-flash | |
Novita AIInference deepseek-v4-flash | $0.14 / $0.28 −67% vs headline | deepseek-v4-flash | |
AIHubMixGateway deepseek-v4-flash | $0.142 / $0.284 −66% vs headline | deepseek-v4-flash | |
PrismCloud deepseek-v4-flash | $0.17 / $0.21 −66% vs headline | deepseek-v4-flash | |
Inference APIsGateway DeepSeek-V4-Flash | $0.19 / $0.38 −55% vs headline | DeepSeek-V4-Flash | |
Alibaba DashScopeLab API deepseek-v4-flash | $0.20 / $0.40 −52% vs headline | deepseek-v4-flash | |
Qianwen AI PlatformLab API deepseek-v4-flash | $0.20 / $0.40 −52% vs headline | deepseek-v4-flash | |
Qwen CloudLab API deepseek-v4-flash | $0.20 / $0.40 −52% vs headline | deepseek-v4-flash | |
CLF AI GatewayCloud deepseek-v4-flash | $0.176 / $0.528 −50% vs headline | deepseek-v4-flash | |
Azure AI FoundryCloud deepseek-v4-flash | $0.19 / $0.51 −49% vs headline Retires 20 Feb 2028 | deepseek-v4-flash | Retires 20 Feb 2028 |
| $0.40 / $0.80 −5% vs headline | deepseek-v4-flash-0731 | ||
LibertAICloud deepseek-v4-flash | $0.25 / $1.75 +19% vs headline | deepseek-v4-flash | |
| $0.424 / $1.27 +21% vs headline | deepseek-v4-flash-0731 | ||
| $0.44 / $1.32 +26% vs headline | deepseek-v4-flash-0731 |
Show 16 more listings: regions, snapshots and variantsHide regional listings
| DeepSeek | |||||||
| Alias deepseek-v4-flash | $0.30 / $1.20 | deepseek-v4-flash | |||||
| DeepInfra | |||||||
| 0731 DeepSeek-V4-Flash-0731 | $0.06 / $0.18 −83% vs headline | DeepSeek-V4-Flash-0731 | |||||
| GMI Cloud | |||||||
| 0731 DeepSeek-V4-Flash-0731 | $0.286 / $0.858 −18% vs headline | DeepSeek-V4-Flash-0731 | |||||
| StreamLake | |||||||
| 0731 deepseek-v4-flash-0731 | $0.44 / $1.32 +26% vs headline | deepseek-v4-flash-0731 | |||||
| Nebius AI Studio | |||||||
| 0731 DeepSeek-V4-Flash-0731 | $0.14 / $0.28 −67% vs headline | DeepSeek-V4-Flash-0731 | |||||
| W&B Inference | |||||||
| 0731 DeepSeek-V4-Flash-0731 | $0.13 / $0.27 −69% vs headline | DeepSeek-V4-Flash-0731 | |||||
| Fireworks AI | |||||||
| Alias deepseek-v4-flash | $0.14 / $0.28 −67% vs headline | deepseek-v4-flash | |||||
| 0731 deepseek-v4-flash-0731 | $0.22 / $0.66 −37% vs headline | deepseek-v4-flash-0731 | |||||
| 0731 deepseek-v4-flash-0731 | $0.22 / $0.66 −37% vs headline | deepseek-v4-flash-0731 | |||||
| Novita AI | |||||||
| 0731 deepseek-v4-flash-0731 | $0.44 / $1.32 +26% vs headline | deepseek-v4-flash-0731 | |||||
| Inference APIs | |||||||
| Alias deepseek-v4-flash | $0.19 / $0.38 −55% vs headline | deepseek-v4-flash | |||||
| 0731 DeepSeek-V4-Flash-0731 | $0.19 / $0.38 −55% vs headline | DeepSeek-V4-Flash-0731 | |||||
| Alibaba DashScope | |||||||
| 0731 deepseek-v4-flash-0731 | $0.20 / $0.40 −52% vs headline | deepseek-v4-flash-0731 | |||||
| Qianwen AI Platform | |||||||
| 0731 deepseek-v4-flash-0731 | $0.20 / $0.40 −52% vs headline | deepseek-v4-flash-0731 | |||||
| Qwen Cloud | |||||||
| 0731 deepseek-v4-flash-0731 | $0.20 / $0.40 −52% vs headline | deepseek-v4-flash-0731 | |||||
| Azure AI Foundry | |||||||
| 0731 DeepSeek-V4-Flash-0731 | $0.44 / $1.32 +26% vs headline Retires 3 Dec 2026 | DeepSeek-V4-Flash-0731 | Retires 3 Dec 2026 | ||||
Capabilities
Core
- Streaming
- Function calling
- JSON mode
- Structured outputs
Multimodal
- Vision
- Image input
- Audio
- Image generation
- Video
Advanced
- Reasoning
- Tool calling
- MCP
- Prompt caching
- Batch API
- Embeddings
- Takes in
- TextImages
- Produces
- Text
- Features
- Assistant prefillParallel tool callsPrompt caching
- Open weights
- License: DeepSeek License (MIT code)Runs in about 698.3 GB of GPU memoryRuntimes: SGLang, TGI, vLLM
Cheaper alternatives
Current models that cost less and keep most of what this one can do: tools, vision, reasoning, structured output, caching and enough context.
Same model, cheaper host
- DeepInfra83% lessDeepSeek-V4-Flash-0731$0.06 / $0.18
- DeepInfra79% lessDeepSeek-V4-Flash$0.09 / $0.18
- Sail79% lessDeepSeek-V4-Flash-0731$0.09 / $0.18
- GMI Cloud78% lessDeepSeek-V4-Flash$0.091 / $0.182
- Pinstripes76% lessdeepseek-v4-flash$0.10 / $0.20
Smaller DeepSeek models
No cheaper DeepSeek model keeps its key capabilities.
Other frontier labs
- 12% less
- 62% less
- 62% less
- 50% less
Open-weight picks
- 8% less
- 17% less
- 33% less
- 51% less
Model line
Newer and older versions, at each lab's own price.
- Sep 2026DeepSeek V4.1 FlashLatest$0.14 / $0.42
- Apr 2026DeepSeek V4 FlashThis page$0.30 / $1.20
Call it
Using DeepSeek's API.
from litellm import completion
response = completion(
model="deepseek/deepseek-v4-flash",
messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)Quick answers
How much does DeepSeek V4 Flash cost?
On DeepSeek, DeepSeek V4 Flash costs $0.30 per million input tokens and $1.20 per million output tokens. Cached input costs $0.006 per million tokens. The cheapest host is DeepInfra (0731), at $0.06 per million input tokens and $0.18 per million output tokens.
Which APIs offer DeepSeek V4 Flash?
DeepSeek V4 Flash runs on 30 APIs: DeepSeek, Sail, DeepInfra, GMI Cloud, Pinstripes, Kyma API, Baseten, Perplexity, ai& and 21 more.
What is DeepSeek V4 Flash's context window?
1M tokens, with up to 393K tokens in one reply.
Is DeepSeek V4 Flash being retired?
Azure AI Foundry (0731 only) retires it on 3 Dec 2026; Azure AI Foundry on 20 Feb 2028. 29 hosts keep serving it.
Is there a newer version of DeepSeek V4 Flash?
Yes. DeepSeek V4.1 Flash is the newest model in this line, released on 10 Sep 2026.
What are cheaper alternatives to DeepSeek V4 Flash?
Models that cost less and keep most of what it can do: GPT-5.4 Nano on OpenAI (12% less), Claude Haiku 5.5 on Anthropic (62% less) and Qwen3.7 Plus on Replicate (8% less). DeepSeek V4 Flash itself costs 83% less on DeepInfra (0731). Savings are on a mix of 3 input tokens to 1 output token.
When was DeepSeek V4 Flash released?
DeepSeek released it on 24 Apr 2026.
What is DeepSeek V4 Flash's API model id?
Use deepseek-v4-flash with DeepSeek. Other hosts use their own ids: DeepSeek-V4-Flash-0731 on Sail, DeepSeek-V4-Flash on DeepInfra, DeepSeek-V4-Flash on GMI Cloud, DeepSeek-V4-Flash-0731 on Baseten.
Sources
- DeepInfra pricing
- GMI Cloud pricing
- Inference APIs pricing
- Nebius AI Studio pricing
- Tensormesh pricing
- W&B Inference pricing
- Azure AI Foundry pricing
- Baseten pricing
- Sail pricing
- Together AI pricing
- Databricks pricing
- ai& pricing
- AIHubMix pricing
- CLF AI Gateway pricing
- Alibaba DashScope pricing
- DeepSeek pricing
- Fireworks AI pricing
- LibertAI pricing
- Novita AI pricing
- Pinstripes pricing
- Prism pricing
- Qianwen AI Platform pricing
- Qwen Cloud pricing
- StreamLake pricing
- Tencent TokenHub pricing
- EmpirioLabs AI pricing
- Perplexity pricing
- Scaleway pricing
- Wally pricing
- LiteLLM model map
Raw catalog entry for deepseek-v4-flash on DeepSeek
{
"cache_creation_input_token_cost": 0,
"cache_read_input_token_cost": 6e-9,
"input_cost_per_token": 3e-7,
"input_cost_per_token_cache_hit": 6e-9,
"litellm_provider": "deepseek",
"max_input_tokens": 1000000,
"max_output_tokens": 393216,
"mode": "chat",
"off_peak_pricing": {
"cache_read_input_token_cost": 3e-9,
"input_cost_per_token": 1.5e-7,
"output_cost_per_token": 6e-7,
"windows": [
{
"hours_utc": [
"00:00-01:00",
"04:00-06:00",
"10:00-00:00"
],
"weekdays": [
1,
2,
3,
4,
5
]
},
{
"hours_utc": "00:00-00:00",
"weekdays": [
6,
7
]
}
]
},
"output_cost_per_token": 0.0000012,
"source": "https://api-docs.deepseek.com/quick_start/pricing",
"supported_endpoints": [
"/v1/chat/completions"
],
"supports_assistant_prefill": true,
"supports_function_calling": true,
"supports_native_streaming": true,
"supports_parallel_function_calling": true,
"supports_prompt_caching": true,
"supports_reasoning": true,
"supports_response_schema": true,
"supports_system_messages": true,
"supports_tool_choice": true,
"supports_vision": true
}