AiCostCompare

DeepSeek V4 Flash

by DeepSeek · deepseek/deepseek-v4-flash · #33 cheapest of 325 paid models

Prices updated Jul 22, 2026, 1:31 AM UTC · refreshed hourly

DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window. It is designed for fast inference and...

Compare vs:Claude Sonnet 5Claude Opus 4.6Claude Haiku 4.5GPT-5.6 Sol

Pricing

Input

$0.094

per 1M tokens

Output

$0.188

per 1M tokens

Blended (3:1)

$0.117

3 input : 1 output

Cache read

$0.019

per 1M cached tokens

Cache write

per 1M tokens

Cache & batch economics

Effective prices for DeepSeek V4 Flash with prompt caching.

Cache hit rateEffective blended $/1MExample request*vs no cache
0%$0.117$0.009755
50%$0.089$0.006003−38%
90%$0.067$0.003002−69%

*Example: 100K sticky context + 2K new input + 1K output. Batch discount is an approximate 50% off list rates (available for OpenAI / Anthropic / Google — not flagged for this creator). Confirm on the vendor’s pricing page.

What a request costs

WorkloadInput tokensOutput tokensCost / requestCost / 1K requests
Short chat message500300$0.000103$0.1032
Document summary8,0001,000$0.000938$0.938
Codebase question (RAG)30,0002,000$0.003189$3.19
Long-context analysis150,0005,000$0.015$15.01

Performance

Intelligence Index

40.3

AA composite quality

Coding Index

56.2

Math Index

Agentic Index

31.1

Output speed

115 tok/s

median

Time to first token

0.86s

median

Time to first answer

49.64s

after reasoning tokens

Artificial Analysis benchmarks

BenchmarkScore
GPQA Diamond89.4%
Humanity's Last Exam32.1%
SciCode44.9%
IFBench79.2%
AA-LCR (Long Context Reasoning)63.0%
Terminal-Bench Hard35.6%
Terminal-Bench 2.161.8%
τ²-Bench (Telecom)95.0%
τ³-Bench Banking22.9%

Design Arena Elo

CategoryElo
Website1,235
UI components1,205
Data viz1,159
SVG1,202
3D1,248
Game dev1,254
Code categories1,239
ASCII art1,157

Specs

Context window

1.05M

tokens

Max output

tokens

Input modalities

text

Tokenizer

DeepSeek