DeepSeek · Open weights

DeepSeek V4.1 Flash Thinking

Added Sep 8, 2026Open weights

DeepSeek V4.1 Flash supports text and image input, reasoning, tool calling, and structured output with a 1M-token context window. This is a rate-limited beta with limited capacity, intended for testing rather than production use.

Reasoning
Tools
Vision
Structured output

Input: Text, Images
Output: Text
Reasoning effort: none · low · high · max
Discovery tier: Open weights

Pricing

USD per 1M tokens

Input: $0.09
Output: $0.25
Cache read: $0.006

Reference prices as published by the model provider. API example

Specifications

Context window: 1M
Max output: 384K
Average output (7 days): 785 tokens; Longer than 64% of 473 models

Side by side

Compare models

Put DeepSeek V4.1 Flash Thinking next to up to three other models. The comparison stays in the link, so you can share it.

TryMuse Spark 1.2 GPT 6 Sol Gemini 3.8 Flash All 3

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

Intelligence Index

39.5

Tested at max reasoning effort

Better than 84% of 326 scored models

3.8Catalog models by score55.1

Agentic work

AutomationBench-AA
68.9%
Workflow automation, penalized for guardrail violations
Better than 97% of 173 scored models
AutomationBench-AA tasks completed
35.5%
Workflows finished end to end without a guardrail violation
Better than 94% of 173 scored models
Harvey LAB-AAv1.1
1.7%
Legal tasks passed in full, with no material hallucination
Better than 75% of 29 scored models
AA-Briefcasev1.1
1,419 Elo
Agentic knowledge work, rated head to head
Better than 75% of 166 scored models
GDPval-AAv2.1
1,600 Elo
Economically valuable occupational tasks, rated head to head
Better than 83% of 187 scored models

Document reasoning

GDP.pdf
12.8%
Professional reasoning over PDF documents, all-pass rate
Better than 56% of 171 scored models
AA-LCRv1.1
84.0%
Reasoning across long documents
Better than 94% of 309 scored models
MLCR-AA
23.9%
Long-context reasoning over medical records
Better than 87% of 121 scored models

Reasoning and knowledge

Humanity’s Last Exam
39.2%
Expert-level questions across academic fields
Better than 74% of 323 scored models

Coding

Terminal-Benchv4.0
26.8%
Practical software engineering in a terminal
Better than 77% of 171 scored models
SciCode
51.9%
Scientific computing in Python
Better than 62% of 174 scored models

Legacy evaluations

AA-LCR (unversioned)
84.0%
Long-context reasoning, earlier grading
Better than 94% of 309 scored models

Scores from Artificial Analysis, updated Oct 26, 2026. DeepSeek V4.1 Flash Thinking was tested at max reasoning effort; compare scores measured at the same effort.Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Throughput: 227.2 tok/s; Faster than 88% of 256 models
Time to first token: 3.7 s; Quicker than 41% of 256 models
Prompt cache hits: 75.8%; About $0.0263 per 1M input tokens with caching
Request success (24h): 99.9%; Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 98.4%

Healthy (98% or more)
Degraded (90–98%)
Down (under 90%)
No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart

curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek/deepseek-v4.1-flash:thinking",
    "reasoning_effort": "none",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'

import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="deepseek/deepseek-v4.1-flash:thinking",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

DeepSeek

Other DeepSeek models, starting with the same family.

Agentic work

AutomationBench-AA

AutomationBench-AA tasks completed

Harvey LAB-AAv1.1

AA-Briefcasev1.1

GDPval-AAv2.1

Document reasoning

GDP.pdf

AA-LCRv1.1

MLCR-AA

Reasoning and knowledge

Humanity’s Last Exam

Coding

Terminal-Benchv4.0

SciCode

Legacy evaluations

AA-LCR (unversioned)