DeepSeek

DeepSeek V4.1 Flash

DeepSeek V4.1 Flash model for reasoning and agentic coding

deepseek/deepseek-v4.1-flash

Context

1M

Output

384K

Released

2026-09-10

Knowledge cutoff

2025-05

Types

InputTextImage
OutputText

Reasoning

Supported

Tool use

Supported

Structured output

Supported

Attachments

Supported

Open weights

Supported

Provider pricing

OpenRouter
InTextImage
OutText
Price
$0.04 / $0.64
Context
1M
Status
Stable
NanoGPT
InTextImage
OutText
Price
$0.13 / $0.52
Context
1M
Status
Stable
CommandCode
InTextImage
OutText
Price
$0.15 / $0.60
Context
1M
Status
Stable
LLM Gateway
InTextImage
OutText
Price
$0.15 / $0.60
Context
1.1M
Status
Stable
Merge Gateway
InTextImage
OutText
Price
$0.15 / $0.60
Context
1M
Status
Stable
ZenMux
InTextImage
OutText
Price
$0.15 / $0.60
Context
1M
Status
Stable
IteraCompute
InText
OutText
Price
$0.29 / $1.16
Context
1M
Status
Stable
CrossModel
InTextImage
OutText
Price
$0.30 / $1.20
Context
1M
Status
Stable
Kilo Gateway
InTextImage
OutText
Price
$0.30 / $1.20
Context
1M
Status
Stable
NEAR AI Cloud
InTextImage
OutText
Price
$0.30 / $1.20
Context
1M
Status
Stable
Novita AI
InTextImage
OutText
Price
$0.30 / $1.20
Context
1M
Status
Stable
Ofox
InTextImage
OutText
Price
$0.30 / $1.20
Context
1M
Status
Stable
Prime Inference
InTextImage
OutText
Price
$0.30 / $1.20
Context
1M
Status
Stable
Vercel AI Gateway
InTextImage
OutText
Price
$0.30 / $1.20
Context
1M
Status
Stable
AI/ML API
InTextImage
OutText
Price
$0.39 / $1.56
Context
1M
Status
Stable
TensorX
InTextImage
OutText
Price
$0.50 / $1.50
Context
1M
Status
Stable

Benchmarks

  • GPQA Diamond

    90.9 · pass@1

  • MathArena Apex

    65.6 · pass@1

  • Humanity's Last Exam

    36.8 · pass@1

  • Humanity's Last Exam

    39.1 · pass@1

  • Codeforces

    3471 · rating

  • Terminal-Bench

    90.6 · pass@1

  • Terminal-Bench

    30 · pass@1

  • Terminal-Bench

    31.2 · pass@1

  • DeepSWE

    74.2 · resolved

  • ProgramBench

    20.3 · almost@1

  • NL2Repo

    64 · score

  • CyberGym

    88.1 · pass@1

  • ExploitGym

    15.3 · pass@1

  • AutomationBench

    54.8 · pass@1

  • Agents' Last Exam

    31.8 · pass@1

  • SEC-Bench Pro

    62.8 · pass@1

  • Humanity's Last Exam

    63.9 · pass@1

  • Chartography

    78.9 · pass@1

  • BabyVision

    89.6 · pass@1

  • ZeroBench

    49 · pass@5

  • DeepSWE

    69.8 · resolved

  • Terminal-Bench

    88 · pass@1

  • DeepSWE

    65.6 · resolved

  • Terminal-Bench

    84.1 · pass@1

  • DeepSWE

    65.5 · resolved

  • Terminal-Bench

    85 · pass@1

  • DeepSWE

    66.2 · resolved

  • Terminal-Bench

    86.1 · pass@1

  • Terminal-Bench

    90.3 · pass@1

  • DeepSWE

    72.6 · resolved

  • DeepSWE

    70.5 · resolved

  • Terminal-Bench

    85.8 · pass@1

  • DeepSWE

    67.6 · resolved

  • Terminal-Bench

    85.8 · pass@1

From the same lab