StepFun Releases Three Large Models of the Step Series

At the WAIC 2024, StepFun launched three new products in the Step series of universal large models: Step-2 trillion-parameter language large model official version, Step-1.5V multimodal large model, and Step-1X image generation large model.

At the same time, StepFun also highlighted self-developed large model application products for C-end users and disclosed the latest developments and plans in the field of large model ecological cooperation.

Among them, the latest release of the Step-2 trillion-parameter language model official version adopts the innovative MoE hybrid expert model architecture. At the same time, based on StepFun’s industry-leading system capabilities, Step-2 has significantly improved training efficiency and comprehensively approaches GPT-4 in areas such as mathematics, logic, programming, knowledge, creativity, and multi-turn dialogue. Enterprises and developers can apply to experience it through StepFun’s open platform.

In addition, StepFun announced the launch of an open platform – “Starry Plan”, investing in high-quality resources to support excellent startups and independent developer teams in the multimodal field, jointly accelerating the application of AI large model technology in various scenarios.

Founder and CEO of StepFun Jiang Daxin said: “Climbing the peak of AGI, ‘trillion parameters’ and ‘multi-modal fusion’ are indispensable. The scale of trillion parameters is the basic threshold for achieving AGI; multi-modal large models are the only way to AGI. In the future, we will continue to make models bigger and stronger, create super models, and at the same time let models serve work and life, ten times everyone’s possibilities.”

SEE ALSO: Tesla’s Optimus Robot to Debut at Shanghai’s 2024 World AI Conference