DeepSeek V4.1 Flash API pricing

DeepSeek V4.1 Flash cost by workload

Effective cost uses the minimum 2× credit.

How DeepSeek V4.1 Flash pricing works

You are charged per token: 0.15 USD per million tokens you send and 0.6 USD per million tokens DeepSeek writes back. Output is 4.0× the input rate for DeepSeek V4.1 Flash, so answer length drives most of the bill.

Somnus bills usage at the list price and credits every top-up at 2× or more, which is how the effective $0.075 / $0.3 rate comes about.

Estimate DeepSeek V4.1 Flash cost from the usage field

from openai import OpenAI
import os

client = OpenAI(
    api_key=os.environ["SOMNUS_API_KEY"],
    base_url="https://somnus.world/api/public/gateway/v1",
)

resp = client.chat.completions.create(
    model="claude-sonnet-5",
    messages=[
      {"role": "user", "content": "Count my tokens"}
    ],
)
print(resp.choices[0].message.content)

Multiply usage.prompt_tokens by 0.15/1e6 and usage.completion_tokens by 0.6/1e6.

Questions

Related