DeepSeek · Very Fast · Updated 2026-11-22

DeepSeek V4.1 Flash API Pricing & Cost Breakdown

Ultra-cheap bilingual CJK/English automation, coding helpers, and high-volume agents.

Input / 1M Tokens
$0.15
Cached Input / 1M
$0.003
Output / 1M Tokens
$0.60
Context Window
1.05M

Workload Cost Scenarios (30-Day Billing)

Off-peak $0.15/$0.60 (peak $0.30/$1.20); cached input drops to $0.003/1M.

Scenario In / Out per Req Uncached Monthly 50% Cache Hit Monthly
Hobby / Prototype (100 req/day) 1500 / 400 $1.40 $1.06
Growing Startup (2,500 req/day) 2500 / 700 $59.63 $45.84
Production Scale (50,000 req/day) 3000 / 800 $1,395.00 $1,064.25

Compare DeepSeek V4.1 Flash Against Alternatives

Lower-Cost Alternative Models

GPT-6 Luna ($0.50/1M out) Claude Haiku 5.5 ($0.50/1M out)

Frequently Asked Questions

What is DeepSeek V4.1 Flash API pricing per 1M tokens?

Standard input is $0.15/1M, cached input is $0.003/1M, and output is $0.60/1M.

What is the context window of DeepSeek V4.1 Flash?

DeepSeek V4.1 Flash supports a 1.05M token context window with up to 393,216 max output tokens.

How much does prompt caching save on DeepSeek V4.1 Flash?

Prompt caching reduces cached input token rates by 98% compared with uncached input.

What workloads is DeepSeek V4.1 Flash best suited for?

Ultra-cheap bilingual CJK/English automation, coding helpers, and high-volume agents.