DeepSeek V4 didn’t explode when released, but it started a revolution by cutting prices?

Original author：老范讲故事

Original source：https://www.youtube.com/embed/4MyGxyLRWJk

Paid channel subscription: https://www.youtube.com/channel/UCUGLhcs3-3y_yhZZsgRzrzw/join DeepSeek V4 didn’t explode when released, but started a revolution with price reduction? After the price reduction of DeepSeek V4, will the Coding Plan fail? The real killer trick of DeepSeek V4 is not the running score, but the pricing? This article breaks down how the continuous price reduction of DeepSeek V4 changes the cost of using AI models and helps developers and low-frequency users decide whether to purchase Coding Plan or Token Plan. The article focuses on DeepSeek V4 price reduction, million token price, cache hit cost, and how to choose between DeepSeek V4 Flash and Pro. Combined with Claude Code's actual experience in writing websites, it explains why Flash is suitable for daily running volume, and Pro is more suitable for complex agents, fact checking, and long-term planning. We will also analyze the subscription logic of Token Plan and Coding Plan, and why low-frequency users may be more suitable for pay-as-you-go. 00:00:00 DeepSeek’s price reduction revolution 00:01:23 The gap in public opinion after the release 00:02:54 Bragging about contributions and experience 00:05:28 Continuous price reductions to break through the package 00:07:22 Rough calculation of the cost of one million Tokens 00:07:59 ClaudeCode actual measurement 00:10:04 Why price reduction is a revolution 00:10:30 How to play the Token package 00:12:20 Hunger marketing is overturned 00:13:02 Xiaomi’s open source and peer pressure 00:14:21 Flash and Pro trade-off 00:16:08 Package logic is broken 00:16:53 DeepSeek’s low-price logic 00:18:29 Open source model grabs the entrance 00:19:39 Self-service package is suitable for people 00:20:40 Welcome to join the Discord discussion server regarding changes in the Guoyun product industry: https://discord.gg/ppKsNkttTv