Skip to content
InferFlux

deepseek-ai/DeepSeek-V4.1-Flash-1017

Aliasdeepseek/deepseek-v4.1-flash-1017by deepseek
Chat with this modelAPI
Opened
Nov 11, 2026
Context
1.05M tokens
Input, per million
$0.275
Output, per million
$1.10
Modalities
text to text
Verification
Open weights. Hugging Face lists this repository with these weight files. Validators check it every day.
Variant
The lab's API
Weights on Hugging Face
deepseek-ai/DeepSeek-V4.1-Flash-1017 at 08c8657weight files 0x9ef7cc22…93ba0b62
Commits that carry these weights
Raw id
0x7582bad3…be1a0a1b

Supported parameters

  • logprobs
  • max_tokens
  • reasoning
  • reasoning_effort
  • response_format
  • stop
  • temperature
  • tool_choice
  • tools
  • top_logprobs
  • top_p

Providers

1 endpoint, cheapest first. Measured by the floor; audited by validators.

One row per endpoint, cheapest first: its provider and where it is served from, its variant, its match rate, prices per million tokens, context, and the floor's measurements
ProviderVariantMatch rateInput /MOutput /MCache read /MContextLatencyThroughputUptimeFloor scoreDetails
InferFlux Provider 1served from provider-1.testnet.inferflux.xyzThe lab's APIapi.deepseek.comdeepseek-flashNo audits yet$0.275$1.10$0.07151.05M–Not yet measured–Not yet measured–Not yet measured–Not yet measured

Pricing

Lowest price

The cheapest endpoint's listed price, per million tokens.

Input
$0.275
Output
$1.10
Cache read
$0.0715

Clearing price

What cleared receipts paid per million tokens over the last day, weighted by tokens, input and output together.

Clearing
$0.88172
Band
$0.390625 to $0.988506
Receipts
412

Performance

Measured by the floor.

Fastest output, median
Not yet measured
Quickest first token, median
Not yet measured
Each provider's measured serving of this model
ProviderThroughput, median95th percentileFirst token, median95th percentileUptime, last dayFloor scoreToken accounting
InferFlux Provider 1–Not yet measured–Not yet measured–Not yet measured–Not yet measured–Not yet measured0.0%–Not yet measured

The floor's own measurements: speed and first token over its recent probes, the floor score over 30 days, token accounting as the share of receipts whose counts held.

API

Any OpenAI-compatible client works with two changes: the base URL and the key. Create a key on the keys page and keep it in INFERFLUX_API_KEY. The quickstart shows the rest.

Base URL
https://testnet.inferflux.xyz/api/v1
Model
deepseek/deepseek-v4.1-flash-1017It also answers to deepseek-ai/DeepSeek-V4.1-Flash-1017, and to its raw id.
curl
curl https://testnet.inferflux.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $INFERFLUX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek/deepseek-v4.1-flash-1017",
    "messages": [{"role": "user", "content": "Hello"}]
  }'
Python
import os
from openai import OpenAI

client = OpenAI(base_url="https://testnet.inferflux.xyz/api/v1", api_key=os.environ["INFERFLUX_API_KEY"])

completion = client.chat.completions.create(
    model="deepseek/deepseek-v4.1-flash-1017",
    messages=[{"role": "user", "content": "Hello"}],
)
print(completion.choices[0].message.content)
TypeScript
import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://testnet.inferflux.xyz/api/v1", apiKey: process.env.INFERFLUX_API_KEY });

const completion = await client.chat.completions.create({
  model: "deepseek/deepseek-v4.1-flash-1017",
  messages: [{ role: "user", content: "Hello" }],
});
console.log(completion.choices[0].message.content);
Parameters this model's endpoints support
NameTypeWhat it does
logprobsbooleanReturns the probability of each token.
max_tokensintegerThe most tokens the answer may hold.
reasoningobjectTurns the model's reasoning on or off, or sets its effort, where it reasons.
reasoning_effort–Passed through to the endpoint as sent.
response_formatobjectRequests JSON, or JSON that matches a schema.
stopstring or arrayText at which the answer ends.
temperaturenumberHow much the sampling varies. Lower is steadier.
tool_choicestring or objectWhether the model must, may or must not call a tool, or which one.
toolsarrayFunctions the model may call.
top_logprobsintegerHow many likely alternatives to return per token.
top_pnumberSamples only from the tokens that make up this share of the probability.