Back to home
DeepSeekReleased 2026-04-30

DeepSeek V4 Flash

DeepSeek V4 Flash is the cheapest pro-quality LLM API in the world at $0.14/$0.28 per 1M tokens. Released April 30, 2026, it offers 1M context window and matches GPT-5.6 quality on most tasks at 18x lower cost.

Context
1M tokens
Input
$0.14/1M
Output
$0.28/1M
Modalities
text

Best For

Cheapest pro-quality APIBulk processingCode completionTranslation

Strengths

  • Cheapest pro-quality API globally ($0.14/$0.28 per 1M)
  • 1M context window
  • Matches GPT-5.6 quality on 80% of tasks
  • Open-source (Apache 2.0)
  • Strong at math and code

Weaknesses

  • Text-only (no multimodal)
  • Less consistent on edge cases
  • Peak pricing: 9:00-12:00 and 14:00-18:00 Beijing time, prices double
  • Chinese-friendly but English slightly weaker than GPT-5.6

Related Alternatives

GPT-5.6 Luna

Similar price, multimodal, English better

Qwen3.8-Max

Open-weights Chinese champion, $2/$6 per 1M

Frequently Asked Questions

How much does DeepSeek V4 Flash cost?

DeepSeek V4 Flash costs $0.14 per 1M input tokens and $0.28 per 1M output tokens during off-peak hours. During Beijing time peak hours (9:00-12:00 and 14:00-18:00), prices double to $0.28/$0.56 per 1M tokens.

Is DeepSeek V4 Flash as good as GPT-5.6?

On most tasks (translation, summarization, code completion, bulk processing), yes — within 5-10% quality. On complex multi-step reasoning and agent workflows, GPT-5.6 still leads by 10-20%. For price-sensitive workloads, DeepSeek V4 Flash is 12-60x cheaper.

See DeepSeek V4 Flash's full record in our database

Live pricing for all DeepSeek plans, token cost calculator, and the latest price changes.

View Full Comparison