DeepSeek V4 Flash

V4 Flash is the model I would use when a job means thousands of calls and the bill matters more than the last bit of quality. It is the smaller sibling of DeepSeek V4 Pro in the DeepSeek V4 family. The two share open weights and the same features, and Flash is the one priced for volume.

You do not need a DeepSeek account: it runs on your Kadraris key and balance in USD, with no subscription. Prompts and outputs are not used to train models.

What you can use it for

• Coding agents that load a big part of the repo before each edit
• Mass tagging, field extraction and summaries of long files
• Support chat for a product with a tight budget per conversation
• Prototypes and load tests that send thousands of requests

Settings and limits

• Context window: 1,000,000 tokens; output up to 384,000 tokens.
• It reasons before answering by default, and those reasoning tokens are charged at the output rate.
• Set max_tokens on short tasks so a simple question does not turn into a long think.
• Streaming, tools for function calling and JSON mode are all supported.
• Same /v1/chat/completions endpoint as every other chat model here.

Example request

from openai import OpenAI

client = OpenAI(
    base_url="https://kadraris.com/v1",
    api_key="sk-kd-...",
)
resp = client.chat.completions.create(
    model="deepseek-v4-flash",
    max_tokens=800,
    messages=[{"role": "user",
               "content": "Return JSON with name, city and "
                          "total from this order text."}],
)
print(resp.choices[0].message.content)

Adjust the fields to your schema and raise max_tokens if answers get cut off.

Questions

What does V4 Flash cost per token?

$0.135 / $0.27 per 1M tokens. Its output rate is the lowest among the chat models on Kadraris, which is why it suits agents and batch work. Failed calls cost $0.

When does V4 Pro make more sense?

Flash is several times cheaper on input and on output, with the same context and output limits. Move a task to DeepSeek V4 Pro only if your own data shows a clear gain. The V4 Flash vs V4 Pro page compares them.

Can I switch to it without rewriting code?

Yes. If you already use the OpenAI SDK, change the base URL to Kadraris and the model name to deepseek-v4-flash.

DeepSeek V4 Flash in your code

from openai import OpenAI

client = OpenAI(base_url="https://kadraris.com/v1", api_key="sk-kd-...")
resp = client.chat.completions.create(
    model="deepseek-v4-flash",
    messages=[{"role": "user", "content": "Suggest three opening shots for a short film about a cafe."}],
)
print(resp.choices[0].message.content)

curl https://kadraris.com/v1/chat/completions \
  -H "Authorization: Bearer $KADRARIS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "deepseek-v4-flash", "messages": [{"role": "user", "content": "Suggest three opening shots for a short film about a cafe."}]}'

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://kadraris.com/v1", apiKey: process.env.KADRARIS_API_KEY });
const resp = await client.chat.completions.create({
  model: "deepseek-v4-flash",
  messages: [{ role: "user", content: "Suggest three opening shots for a short film about a cafe." }],
});
console.log(resp.choices[0].message.content);