GuShiio.com Jiang Daxin said that the development of agents relies on two key elements: one is multimodal capabilities and the other is reasoning capabilities.
On February 21, Step Up Star held the Step UP Ecological Open Day. This big model company, which rarely takes the initiative to speak out, has showcased a large number of scenarios where its Step series big models are used in the industry.
After DeepSeek swept the global AI industry chain at the beginning of the year, in addition to discussing its technological innovation, a long-established industry consensus was reiterated, namely,”two-wheel drive” of models and applications.
The technical capabilities of large model companies need to be applied to undertake and demonstrate, which is the basis for the widely recognized performance of their models and a prerequisite for creating a commercial closed-loop.
Agents are currently the most important technology and application direction of Step Star. Jiang Daxin, founder and CEO of Step Star, said at the scene that the company has currently entered the development stage of Agent on the AGI roadmap.
Jiang Daxin said that the development of agents relies on two key elements: one is the multimodal ability, which allows agents to fully perceive and understand the world; the other is the reasoning ability, which enables agents to conduct slow thinking with long thinking chains and proactively plan, try, reflect, and provide accurate answers through constant error correction.
Aiming at these two directions, the Step Step series model matrix has released 11 multimodal large models, covering comprehensive capabilities in speech recognition, speech generation, multimodal understanding, image and video generation. In January this year, Step Star released Step R-mini, becoming the first inference model in the Step series, further expanding the capabilities boundaries of large models.
On this basis, Step Star has made a large number of industrial applications, involving finance and finance, content creation, new retail, digital people and other fields.
Taking the field of finance and economics as an example, Step Star’s multimodal and reasoning capabilities have been deeply applied here. Last year, Step Star cooperated with Interface Financial Union to establish the large model technology company Caiyue Star. Its business focuses on the application of large models in the financial and financial fields.
On the model side, Caiyue Xingchen launched the Caiyue Large Model for the financial industry, and was awarded the “Financial Field Large Model Evaluation Report (2024)” released by Shanghai Artificial Intelligence Laboratory in conjunction with Shanghai University of Finance and Economics and Shanghai Coupus Technology Co., Ltd. Modality first.
On the product side, the company launched the “AI Little God of Wealth” intelligent assistant product in the financial field, which can provide users with functions including AI data mining, AI dialogue and financial report interpretation. In addition, Step Star also teamed up with Guotai Junan and Interface Financial Union to launch the “Junhong Lingxi Model”, a multi-modal securities vertical model with hundreds of billions of parameters, taking the lead in adding large model capabilities to the user’s intelligent service system.
“AI Little God of Wealth” is currently online with Weixin Mini Programs and is in normal operation. In terms of multimodal capabilities, its latest iteration adds financial data charting functions. When users ask questions involving market prices, funds, financial reports and other data, the answers can generate graphical charts that allow investors to quickly find the key points and quickly interpret them.
At the same time, compared with similar products, its core advantages lie in two points: one is that it has in-depth reasoning capabilities, and the other is that it has higher-quality credible corpus. The combination of the two can generate more in-depth and accurate market information. interpretation.
According to interface journalists, the future iteration target of this product is also an agent in the financial field, and will become a social service product with a larger user scale.
“AI Little God of Wealth” is the core display of step-by-large model capabilities in the financial field. This product will also show more unique advantages under the flywheel of “data + scenario”. This may be an important implementation paradigm for large model companies in the industrial field in the future.
