On the morning of August 8th Beijing time, OpenAI released GPT-5. This generation of product integrates large language models with reasoning models, capable of intelligently scheduling multiple sub-models and calling the optimal engine when solving different tasks. Compared to GPT-4o, its factual error rate has also significantly decreased.
Although some industry professionals evaluate it as lacking disruptiveness, its improvements in fact accuracy, multi-round reasoning, and complex task execution still temporarily maintain its position at the global top.
[The rest of the translation follows the same professional and accurate approach, maintaining the original structure and meaning while translating into English.]Without new funding, the scale of investment and model iteration speed will be forced to slow down, forming a vicious cycle of "financing obstruction - innovation slowdown - user loss".
According to "Love Analysis" research statistics, AI large model projects in 2024 have a cumulative project amount exceeding 40 billion yuan. Among them, iFlytek (002230.SZ), Baidu, Alibaba, Tencent, and other major companies have received billion-yuan orders. However, Moonshot currently has no publicly known benchmark orders.
Image source: Love Analysis
Industry insiders feedback that after experiencing the rise and subsequent stagnation of blockbuster models like DeepSeek, AI large model investors have become cautious, preferring to concentrate bets on the top "first two". The survival space for players below the middle tier continues to be compressed.
In internet business history, there is no shortage of companies that have made a sudden splash, but what is lacking is the ability to turn that initial splash into a long-term resonance.
This article is from the WeChat public account "Alpha Factory Research Institute", author: Gojo Satoru, published with authorization by 36kr.
