Private AI

Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Search
 
 
 
 
All conversations

Create

Chat
Studio

Explore

Models
Pricing
AI Apps
Blog

Library

Gallery
Workspace
Usage & billing
API
Suggestions and bugs
Use cases

|

Terms

|

Privacy

|

Refunds
Usage & billing
Teams
Referrals
Presets
Settings
FAQPress & mediaBrand assetsBenchmarksEarn crypto
  1. Models
  2. Text
  3. DeepSeek

DeepSeek V4 Flash

DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with a 1M-token context window, built for fast inference, high-throughput workloads, reasoning, coding, and agent workflows.

  • Reasoning
  • Tool Calling
  • Structured Output

Added Apr 24, 2026

Model weights

Pricing

Auto routing · per 1M tokens
Input
$0.14
Output
$0.28
Cache read
$0.0028
Compare provider prices

Specifications

Context window
1M
Max output
384K
Parameters
284B / 13B
Total / active
Avg output (7d)
142 tokens
Longer than 8% of models

Benchmarks

Sourced from LMArena.

Arena Score

1435.7

Overall Rank

#104 / 419

Votes

51,862

Confidence Interval

1431.7 - 1439.7

Category Scores

Coding

#107 / 414

15,283 votes

1484.1

Math

#125 / 404

2,700 votes

1423.5

Longer Query

#95 / 397

23,298 votes

1447.7

Creative Writing

#96 / 417

8,512 votes

1407.9

Instruction Following

#101 / 419

17,906 votes

1428.1

Hard Prompts

#101 / 419

34,688 votes

1459.0

Additional Categories
22

Multi Turn

#83 / 417

9,343 votes

1452.9

Korean

#92 / 290

902 votes

1393.7

Industry Legal And Government

#96 / 392

4,130 votes

1448.6

Japanese

#96 / 288

650 votes

1397.1

Polish

#97 / 244

1,029 votes

1436.4

Industry Medicine And Healthcare

#98 / 390

3,660 votes

1458.9

Industry Life And Physical And Social Science

#99 / 417

8,380 votes

1456.3

Industry Writing And Literature And Language

#100 / 418

12,547 votes

1415.7

Spanish

#100 / 302

1,576 votes

1431.9

Industry Entertainment And Sports And Media

#102 / 417

11,174 votes

1404.5

Hard Prompts English

#104 / 417

16,193 votes

1463.5

Russian

#104 / 383

5,720 votes

1432.2

Exclude Ties

#105 / 419

39,217 votes

1431.5

Non English

#105 / 419

28,787 votes

1422.8

English

#108 / 419

23,074 votes

1444.9

Chinese

#109 / 395

2,986 votes

1470.1

Industry Business And Management And Financial Operations

#109 / 412

10,418 votes

1438.9

French

#110 / 301

1,998 votes

1449.3

Industry Software And It Services

#110 / 419

21,359 votes

1470.9

Expert

#111 / 369

5,394 votes

1457.2

German

#112 / 320

839 votes

1424.7

Industry Mathematical

#119 / 402

2,974 votes

1433.9

Published 2026-11-06 · Matched as DeepSeek-V4-Flash

LMArena Dataset

Providers

Choose or restrict providers with the API. Provider-selection prices include the selection markup. Provider selection guide · Pricing API (JSON, USD per 1K tokens)

Auto
Auto (ZDR)
$0.14$0.28$0.0028N/AN/A
Auto
Auto (Anonymized)
$0.14$0.28$0.0028N/AN/A
Relace
Relace
FP4
UNK
$0.0279$0.472$0.02790.5s106.5 tps
Runware
Runware
10% off
UNK
$0.0798$0.0718$0.161$0.145$0.0147$0.0132N/AN/A
StreamLake
StreamLake
FP8
CN
$0.0882$0.176$0.01764.1s18 tps
DeepInfra
DeepInfra
FP8
US
$0.0945$0.189$0.01891s32 tps
Venice
Venice
US
$0.101$0.202$0.02061.5s22 tps
Fireworks
Fireworks
US
$0.11$0.221$0.02210.7s72 tps
Uomi
Uomi
GLB
$0.13$0.295N/AN/AN/A
Wafer
Wafer
MXFP4
US
$0.136$0.242$0.02940.4s78 tps
NeuralWatt
NeuralWatt
US
$0.147$0.294$0.0294N/AN/A
AtlasCloud
AtlasCloud
FP4
US
$0.147$0.294$0.02944.4s30 tps
Baidu
Baidu
FP8
CN
$0.147$0.294$0.02943.3s48 tps
Parasail
Parasail
FP8
US
$0.147$0.294$0.07350.6s111 tps
Mancer
Mancer
FP8
$0.157$0.525$0.07871.1s33 tps
Privacy:ZDR, prompts not retainedNo training on promptsRetention and training unclearBest latency and throughput are highlighted.
Batch API
ProviderInputOutputCached input
Auto
Autodeepseek/deepseek-v4-flash
$0.14/M$0.28/M$0.0028/M

Batch rows for this model use normal routing and cost the same as standard API requests. Set provider in a row to choose a provider.

Related text models

Compare DeepSeek V4 Flash with similar models from the same provider or model family.

DeepSeek V4.1 Flash 0910 Uncensored

deepseek/deepseek-v4.1-flash-0910-uncensored

Uncensored DeepSeek V4.1 Flash with text and image input, optional reasoning, and tool calling. The date label is retained for compatibility; the served checkpoint can vary with availability.

DeepSeek V4.1 Flash 0910 Uncensored Thinking

deepseek/deepseek-v4.1-flash-0910-uncensored:thinking

Uncensored DeepSeek V4.1 Flash with text and image input, optional reasoning, and tool calling. The date label is retained for compatibility; the served checkpoint can vary with availability.

DeepSeek V4.1 Flash 1017 Uncensored

deepseek/deepseek-v4.1-flash-1017-uncensored

Uncensored DeepSeek V4.1 Flash with text and image input, optional reasoning, and tool calling. The date label is retained for compatibility; the served checkpoint can vary with availability.

DeepSeek V4.1 Flash 1017 Uncensored Thinking

deepseek/deepseek-v4.1-flash-1017-uncensored:thinking

Uncensored DeepSeek V4.1 Flash with text and image input, optional reasoning, and tool calling. The date label is retained for compatibility; the served checkpoint can vary with availability.

DeepSeek V4.1 Flash 1017

deepseek/deepseek-v4.1-flash-1017

The October 17, 2026 snapshot of DeepSeek V4.1 Flash supports text and image input, tool calling, and optional reasoning with low, high, and max effort. It has a shared 1,048,576-token context window and supports up to 393,216 output tokens within the remaining context. Released under the MIT license.

DeepSeek V4.1 Flash 1017 Thinking

deepseek/deepseek-v4.1-flash-1017:thinking

The October 17, 2026 snapshot of DeepSeek V4.1 Flash supports text and image input, tool calling, and optional reasoning with low, high, and max effort. It has a shared 1,048,576-token context window and supports up to 393,216 output tokens within the remaining context. Released under the MIT license.