OpenAI"断供",国产大模型"抢客"

The Paper

2024-06-26 14:04Posted on the official account of Shanghai The Paper

OpenAI"断供",国产大模型"抢客"busy。

On June 25, artificial intelligence (AI) giant OpenAI informed API (Application Interface) users through an official push email that from July 9, API traffic from regions not included in the list of supported countries and regions will be blocked, and if you want to continue to use OpenAI's services, you need to access them in supported regions.

OpenAI API通知邮件

OpenAI announced in March 2023 that it would open APIs that allow third-party developers to integrate ChatGPT into their applications and services through APIs. Currently, a total of 188 countries and regions support access to API services, and accessing OpenAI API services outside of supported regions may result in account bans.

This also means that OpenAI will also stop China's API services that are not on the list.

OpenAI's move is believed to be the result of a combination of factors such as legal compliance.

Announcement of supported countries and regions for OpenAI's official website API

OpenAI "cut off supply", and domestic large-scale model manufacturers immediately began to "grab customers". As of press time, according to incomplete statistics from reporters, nearly ten domestic large models have announced the provision of migration services.

Baidu has launched the "Hometown Cloud and Domestic Large Model Inclusive Program" to provide zero-cost switching services for newly registered enterprise users, and to give OpenAI migrated users an additional ERNIE3.5 flagship model Tokens package with the same scale as OpenAI.

Alibaba Cloud announced that it will provide OpenAI API users with a more cost-effective alternative to China's large models, and announced that Chinese developers will provide 22 million free tokens and exclusive migration services.

Zhipu AI announced the official launch of the OpenAI API user special relocation plan, providing developers with 150 million tokens (50 million GLM-4 + 100 million GLM-4-Air) and a series of migration training from OpenAI to GLM. Zhipu also said that it will provide high-usage customers with an unlimited Token gift program equal to OpenAI's usage scale, as well as a concurrency scale equivalent to OpenAI.

MINIMAX has also launched a "zero-cost nanny migration solution", which provides platform interfaces to adapt to OpenAI, and can use large models of MOE architecture for free. 010000 launched the "Yi API Two-Fold Replacement Plan" and launched a smooth migration service for OpenAI users to Yi series large models. Tiangong Open Platform announced that it will provide 150 million Tokens of API call resources, and will also provide technical support for Tiangong migration services to developers. Baichuan model announced that it will give away 10 million tokens, and the Assistants API is free to use. Silicon-based Flow Announces Silicon Cloud Opens 7 Large Models for Free.

At the same time, the dark side of the moon also issued an article again emphasizing that the Kimi open platform interface is fully compatible with OpenAI, and the fastest five minutes can achieve a silky "move". When the Moonshot AI Open Platform launched its public beta in February this year, the Dark Side of the Moon said that in order to facilitate AI developers to quickly access model capabilities, the API of the open platform is compatible with OpenAI. Developers can migrate smoothly without having to make any "additional" changes to the code beyond the basic parameters.

The supply interruption of OpenAI will affect domestic companies and individuals using its services to a greater or lesser extent, and the development of domestic large models may begin to accelerate as a result.

Han Sirui, an assistant professor at the Hong Kong University of Science and Technology and senior business development manager of the Hong Kong Generative Artificial Intelligence R&D Center, told The Paper that this incident will promote the increase in the market share of domestic large-scale model companies in the short term; On the other hand, in terms of model Chinese capabilities, many domestic models actually perform better than OpenAI, and the improvement of capabilities has also increased the number of users of domestic models themselves.

Han Sirui also said frankly that OpenAI's model capabilities are still relatively leading in advanced tasks such as complex document analysis and complex document generation in text models, and the leading position in technology must be recognized, so the demand for these complex model capabilities will not change greatly in the short term because the API stops serving.

Regarding the preferential migration of various large models, an industry insider told the surging news reporter, "In fact, everyone's API is aligned with OpenAI, and the price is also advantageous compared with OpenAI", which is just a suitable node this time.

Zhou Hongyi also posted a video on Weibo saying that OpenAI's suspension of service in China can only accelerate the development of China's own large model industry, which is not necessarily a bad thing.

He explained: "OpenAI's API cannot be called, which forces domestic applications to choose only domestic large models, and the gap between domestic large models and GPT has gradually narrowed...... This has no impact on enterprise applications and entrepreneurs, and they can deploy local large models on their own. Large models will gradually be migrated locally, which is faster and more secure. ”

View original image 109K

• 

•