DeepSeek V4 Launch Day Reactions from Chinese and American Developers: Architectural Innovation Recognized, but Consensus Remains "Still Half a Year Behind the Frontier"

According to Beating Monitoring, after the release of DeepSeek V4, the Chinese and American developer communities show a coexistence of “technological recognition and gap consensus.”
In the United States, Replit CEO Amjad Masad praised V4’s attention compression and improved efficiency for long context as genuine architectural innovation, and developers on Reddit and Hacker News responded positively to the 1M context open-source model and MIT license, but most adopted a wait-and-see attitude of “test first, then decide.”
CFR researcher Chris McGuire pointed out that V4 reports itself admits to being 3 to 6 months behind cutting-edge models.

In China, V2EX and Zhihu communities focus on V4’s programming agent capabilities and low-cost strategy, with early feedback suggesting the Pro version is close to Claude Opus 4.6 level, but still lags in complex deep reasoning tasks.
There is high attention to Huawei Ascend adaptation. Several months before release, the community had accumulated many doubts due to repeated delays, and after the official release, it shifted into intensive testing.
Both sides share a common concern: V4 performs strongly in coding and long context scenarios, but in comprehensive reasoning, it still has not caught up with the leading closed-source models in the United States.