deepseek-v4-flash

Models/deepseek-v4-flash

Model Code

deepseek-v4-flash

Model overview

Text generationDeep thinking

DeepSeek-V4-Flash is an efficient lightweight MoE model with 284B total parameters and 13B active parameters. It natively supports million-token long context, offers fast inference and low latency, and is suited for high-concurrency lightweight tasks such as daily conversation, content creation, basic RAG, and batch copywriting.

Company: DeepSeekProvider: Alibaba Cloud

Model capabilities

Input modalityT

Output modalityT

Model experience✓

Function calling✓

Structured output-

Web search-

Prefix completion-

Context cache✓

Batch inference✓

Fine-tuning-

Model pricing

Input$0.1493/1M tokens

Input (cache hit)$0.02985/1M tokens

Output$0.2985/1M tokens

Rate limits and context

Maximum input length1M

Context length1M

Maximum output length384K

RPM15000

TPM1200000

API reference

EndpointPOST /v1/chat/completions

SDKOpenAI-compatible

API code examples

Get API Key

from openai import OpenAI

client = OpenAI(
    api_key="$WLROUTER_API_KEY",
    base_url="https://api.wlrouter.com/v1"
)

response = client.chat.completions.create(
    model="deepseek-v4-flash",
    messages=[
        {"role": "user", "content": "Reply with one short sentence."}
    ],
    max_tokens=32
)

print(response.choices[0].message.content)