AI BriefingDeepSeek makes 75% discount permanent
·2026.05.23 08:28
Key point
DeepSeek announced that it is permanently applying a 75% discount to v4 Pro model pricing.
Details
DeepSeek has turned the 1/4 discounted pricing for the v4 Pro model into a permanent policy.
New permanent pricing (75% discount applied):
- Input tokens per 1M (cache miss): $0.435 (previously $1.74)
- Output tokens per 1M: $0.87 (previously $3.48)
- Input tokens per 1M (cache hit): $0.003625 (previously $0.0145)
This makes it 12-17x cheaper than frontier models like Sonnet 4.5 and GPT, and 60-80x cheaper on cache hits. DeepSeek models are known to have high cache hit rates, making the actual cost savings even greater.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.