Back to the models list Active on this gateway

Deepseek: deepseek-v4-flash

deepseek/deepseek-v4-flash Start building free

deepseek-v4-flash is a variant of DeepSeek’s fourth-generation large language model, provided in a "flash" tier that emphasizes very low latency and cost for interactive applications. It belongs to the DeepSeek-V4 family and is tuned to deliver fast general-purpose chat, coding assistance, and reasoning with a more lightweight serving profile than the flagship, heavier V4 models. This tier is typically intended for high-throughput, user-facing workloads such as chatbots, assistants, and simple agents where response speed and efficiency matter more than absolute peak accuracy. It trades some raw capability for faster responses and more economical deployment, while still supporting the broad, multilingual and coding-focused capabilities expected from modern LLMs.

Modalities
Context 1M tokens
Quota cost Standard (1×)
Added September 23, 2026

Pricing

Current price per 1,000,000 tokens. Every request is billed automatically from its exact usage.

Free to use Included at no cost — no input or output charges for this model.

Included in plans

This model is not included in any plan yet.

Statistics

Live performance and usage of this model, aggregated across all users.

Rolling 30 days Aggregated across all users · refreshed every few minutes
Requests 269 ▲ 13,350% vs previous 30 days
Success rate 99,25% 2 failed requests
Median latency 3,86 s p95 33,49 s
Throughput 73,6 output tokens per second
Tokens processed 3M 11,211 avg. per request
Est. cost per request Free no usage charges
Popularity #4 of 35 1,27% of all requests
Uptime (24h) 96,36% 57 requests

Usage over time

Token usage

Last 30 days
Input tokens
2,9M 94,7%
Output tokens
159,8K 5,3%
Avg. input per request
10,617
Avg. output per request
594
Largest request
173,573 tokens

Reliability

Last 30 days
Successful
264 98,1%
Failed
2 0,7%
No response
3
  • 200 Success 264 98,1%
  • — No response 3 1,1%
  • 400 Bad request 2 0,7%

Response time

Successful requests · last 30 days
Fastest353 ms
Average8,23 s
Median3,86 s
95th percentile33,49 s
99th percentile53,58 s
Slowest79,41 s

Last 24 hours

≥ 99% ≥ 90% < 90%
Latency (ms)
Requests per hour

Usage rhythm

Last 30 days · site timezone
  • Busiest hour: 12:00
  • Busiest weekday: Saturday
  • Busiest day: 31 October (89 requests)

All-time record

Total requests
271
Total tokens
3M
Active days
15
First request
23 October 2026
Last request
15 ساعت ago

API

Call this model from any OpenAI-compatible client using your API token.

curl
curl https://apiex.cloud/api/v1/chat/completions \
  -H "Authorization: Bearer $NEW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "deepseek/deepseek-v4-flash",
  "messages": [
    {
      "role": "user",
      "content": "Explain quantum entanglement in one paragraph."
    }
  ],
  "temperature": 0.7
}'
Authentication

Every request must include an Authorization: Bearer <TOKEN> header. Generate a token from your account's API Keys page — its access to models is set automatically from your active plans and redeemed codes, and you can set its rate limit and IP restrictions.

Supported parameters
Parameter Type Description
modelstring, requiredThis model's ID, exactly as shown above.
messagesarray, requiredConversation history as role/content objects (system, user, assistant).
temperaturenumber, optionalSampling randomness, typically 0–2.
max_tokensinteger, optionalUpper limit on tokens generated in the reply.
top_pnumber, optionalNucleus-sampling threshold, as an alternative to temperature.
streamboolean, ignoredResponses are always returned in full so the exact cost can be billed; streaming is not available.

Replace YOUR_API_KEY with a token from your account, then see the full API reference for every parameter.

Frequently asked questions

What is deepseek-v4-flash?

deepseek-v4-flash is a variant of DeepSeek’s fourth-generation large language model, provided in a "flash" tier that emphasizes very low latency and cost for interactive applications. It belongs to the DeepSeek-V4 family and is tuned to deliver fast general-purpose chat, coding assistance, and reasoning with a more lightweight serving profile than the flagship, heavier V4 models. This tier is typically intended for high-throughput, user-facing workloads such as chatbots, assistants, and simple agents where response speed and efficiency matter more than absolute peak accuracy. It trades some raw capability for faster responses and more economical deployment, while still supporting the broad, multilingual and coding-focused capabilities expected from modern LLMs.

How much does deepseek-v4-flash cost?

Input tokens are priced at $0,1092 per 1,000,000 tokens and output tokens at $0,2184 per 1,000,000 tokens. Pricing is billed automatically per request; no separate plan is required.

What is the context length of deepseek-v4-flash?

deepseek-v4-flash supports a context window of up to 1,048,576 tokens.

Does deepseek-v4-flash support image input?

deepseek-v4-flash currently accepts text input only.

Which provider is deepseek-v4-flash from?

deepseek-v4-flash is developed by Deepseek and made available on this gateway through our routing infrastructure.

When was deepseek-v4-flash added to this gateway?

deepseek-v4-flash has been available on this gateway since September 23, 2026.

Is deepseek-v4-flash currently available?

Yes, deepseek-v4-flash is active and can be used right now through this gateway.

Explore more models

Keep exploring