The article mainly introduces a major initiative recently announced by DeepSeek: a permanent reduction in the API price of DeepSeek-V4-Pro models.
Article author and source: Li Yuxia
On May 22, DeepSeek officially announced a major piece of news: the API price of the DeepSeek-V4-Pro model will be permanently and significantly reduced , turning the original "limited-time offer" into a "permanent price reduction"!
The new rates will officially take effect on May 31, 2026 , with the adjusted price being 1/4 of the original price (i.e., 25% off) , setting a new low for global large-scale models.
What does this price reduction mean?
For regular users:
If you usually only use DeepSeek's official website or app for chatting, it remains completely free and is unaffected by this API price adjustment.
For developers and businesses:
This is a truly significant benefit. Especially for groups that need to make a large number of API calls to write code, create intelligent agents, or handle scenarios with high cache hit rates (such as the RAG knowledge base), the cost reduction exceeds 90%, with some netizens exclaiming that "token freedom has been achieved."
Industry comparison:
In the international market, this fee rate is extremely competitive. For example, OpenAI's recently released GPT-5.5 output is priced at approximately $30 per million tokens (about RMB 216), which is more than 30 times more expensive than DeepSeek.
Why is DeepSeek willing to lower its prices against the trend?
At a time when many large model manufacturers in the industry have chosen to raise prices or charge fees, DeepSeek dares to offer prices at rock-bottom levels, mainly due to two factors:
Technological breakthroughs reduce costs:
The new generation V4-Pro model, through architectural innovations (such as the hybrid attention mechanism), reduces the computational power consumption per inference to 27% of the previous generation, and the memory usage is only one-tenth of the original. This significant cost reduction makes the permanent price reduction commercially viable.
Supported by substantial financing:
At the same time as announcing the price reduction, DeepSeek is pushing forward with a funding round of up to 70 billion yuan, with founder Liang Wenfeng even planning to personally invest about 20 billion yuan. This indicates that the company's current strategic focus is on prioritizing the expansion of technological boundaries and seizing the high ground in AGI (Artificial General Intelligence), rather than rushing for short-term monetization.
In short, DeepSeek is using extremely low prices to attract global developers and accelerate the building of its AI ecosystem. If you or your friends need access to large model APIs, now is indeed a very worthwhile opportunity to get started!
Data and images sourced from: DeepSeek's official website, publicly available online information, etc.
