Back to home
DeepSeekReleased 2026-05-31

DeepSeek V4 Pro

DeepSeek V4 Pro is the higher-quality sibling of V4 Flash at $0.435/$0.87 per 1M tokens. After a May 2026 promotional period, the price became permanent. About 3x more expensive than V4 Flash but 30% better quality.

Context
1M tokens
Input
$0.435/1M
Output
$0.87/1M
Modalities
text

Best For

Higher quality than V4 FlashComplex reasoningProduction workloads

Strengths

  • Higher quality than V4 Flash on hard tasks
  • Still ~6x cheaper than GPT-5.6 standard
  • 1M context window
  • Open-source (Apache 2.0)
  • Strong at math and code

Weaknesses

  • Still text-only
  • Cache hit pricing varies across vendors (SiliconFlow $1.0/1M vs DeepSeek official $0.025/1M — 40x spread)
  • Peak hour pricing applies
  • Older name 'deepseek-chat' retires July 24, 2026

Related Alternatives

DeepSeek V4 Flash

Cheaper, similar quality for most tasks

Qwen3.8-Max

Larger 2.4T params, $2/$6 per 1M

Frequently Asked Questions

How much does DeepSeek V4 Pro cost?

DeepSeek V4 Pro costs $0.435 per 1M input tokens and $0.87 per 1M output tokens — permanently (the price after the May 2026 promo became permanent). The old name 'deepseek-chat' retires July 24, 2026.

Is V4 Pro worth 3x the price of V4 Flash?

For most tasks, no — V4 Flash gives 90% of V4 Pro quality at 1/3 the price. V4 Pro is worth the premium for complex reasoning, agent workflows, and cases where quality matters more than cost.

See DeepSeek V4 Pro's full record in our database

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

View Full Comparison