Wen Xin Yiyan is upgraded and Wen Xin Qianfan is released! Is MaaS starting to roll in?

On March 27, at a closed-door communication meeting held by Baidu Intelligent Cloud, Baidu officially launched the "Wenxin Qianfan" large model platform.

It is reported that this is an enterprise-level large language model service platform provided by Baidu to customers. The Wenxin Qianfan large model platform includes not only Wenxin Yiyan, but also Baidu's full set of Wenxin large models and corresponding development tool chains. In the future, Wenxin Qianfan will also support third-party open source large models.

Wen Xin Qianfan: Baidu’s first step in promoting the commercial application of AI large models

At the communication meeting, Baidu conducted a real-time demonstration, covering corporate offices, e-commerce There are five major application scenarios: business services, tourism services, government offices, and financial services.

In corporate offices, Wenxin Qianfan can make PPT in 3 minutes; in e-commerce services, Wenxin Qianfan can quickly generate product promotion words and initiate digital live broadcasts to bring goods; In travel services, Wen Xin Qianfan can generate travel plans and complete air ticket and hotel reservations after simple communication.

Previously, Baidu CEO Robin Li mentioned that Wen Xinyiyan will provide external services through Baidu Intelligent Cloud to help companies build their own models and applications. Now it seems that this idea is quickly coming to fruition.

In terms of charging model, Wen Xin Qianfan’s inference service calls are charged at 0.012 yuan/1000 tokens, and payment is based on the total number of input and output words of the call. Previously, OpenAI officially announced the opening of API, and the interface service is priced at US$0.002 (≈ RMB 0.014 yuan)/1000 tokens.

Judging from the demonstration, Wen Xin Qianfan has opened up the imagination of specific applications for industry users; from the perspective of charging, it is basically within a reasonable range. As for the actual commercial effect, it still requires further evaluation and feedback from companies that are brave enough to try it.

From IaaS to MaaS, how far is it?

At the Wen Xin Yiyan press conference held in the middle of this month, Robin Li said that the era of large models will produce three Big industry opportunities.

The first category is a new type of cloud computing company, whose mainstream business model has changed from IaaS to MaaS.

The second category is companies that fine-tune industry models. This is the middle layer between general large models and enterprises, providing solutions for industry customers.

The third category is companies that develop applications based on large model bases, that is, application service providers.

The MaaS mentioned here is Model as a Service, model as a service. Concepts similar to SaaS (Software as a Service) and IaaS (Infrastructure as a Service) refer to a certain business model.

MaaS is not a particularly new concept, but it wasn’t until large model products like ChatGPT became popular around the world that MaaS gradually entered people’s horizons. The business of “raising” models was very fast. Become a trend that continues to attract attention.

Simply put, the MaaS model refers to the provision of pre-trained models by capable large companies, allowing small companies in vertical industries to build and deploy AI models without investing in building and The infrastructure, hardware, and expertise required to maintain your own models.

It is foreseeable that in the future, more and more AI model developers will provide service subscribers with solutions for specific scenarios, and the solutions themselves are executed by the AI ​​model (service )of. At present, the prospects of this business model are quite broad, especially when multi-modal large-scale models are further developed, AI implementation scenarios are expected to achieve leapfrog growth.

ChatGPT’s accelerated commercialization is both pressure and motivation

Since the release of ChatGPT, the speed at which its team has updated AI products and technologies has been overwhelming. The emergence of ChatGPT is like holding a torch in the dark night, which not only ignites public enthusiasm, but also illuminates the way forward for the commercialization of large AI models.

Whether it is combined with Bing Browser or upgraded to an online version, ChatGPT is creating a new Internet portal driven by artificial intelligence. The acceleration of its commercialization is obvious to all, which will inevitably put pressure on many companies in the same ecosystem, but it also opens up a ladder of advancement for them.

Due to long-term business and security needs, major domestic companies such as Baidu must continue to develop and optimize their own large models. Driven by the Maas business model, leaders with large model foundations will one day become AI infrastructure providers and community ecosystem builders.

Reference link:

​​//m.sbmmt.com/link/ 45f6a4a57549a5720dfdcdf643c78b83##​​

​##//m.sbmmt.com/link/4ef30115b941a1bcd71ffebb50c69ece​​​

//m.sbmmt.com/link/dfc7defac6624a80f02b02e22b14e8fd​​

The above is the detailed content of Wen Xin Yiyan is upgraded and Wen Xin Qianfan is released! Is MaaS starting to roll in?. For more information, please follow other related articles on the PHP Chinese website!