SciMaster

The world's first general-purpose scientific research AI agent, jointly released by Shanghai Jiao Tong University and Shenzhen Technology

academic research

https://scimaster.bohrium.com/

Visit Website

In-depth Report

SciMaster is the world's first general-purpose scientific research AI agent jointly released by Shanghai Jiao Tong University, Shenzhen Technology and Shanghai Algorithm Innovation Institute in July 2025. It represents an important milestone in the transformation of the field of AI for Science from "revolution of tools" to "revolutionary tools". This product is built on the large model of the Innovator scientific base and can compress scientific research tasks that traditionally take several days to be completed within 30 minutes. The core concept of SciMaster is to make scientific research no longer just for scientists, to enable more people to participate in scientific research, and to promote the scientific research paradigm towards "universal scientific intelligence". The product is currently in the internal testing stage, and basic functions are free to use. Users can apply for an invitation code through the official website to experience it.
SciMaster was born from the rapid development of AI for Science technology. On July 26, 2025, Shanghai Jiao Tong University, Shenzhen Technology and Shanghai Algorithm Innovation Institute jointly released the Innovator, a truly universal scientific base model, and relied on this model to launch the world's first universal scientific research intelligent agent, SciMaster. This achievement marks the transformation of the role of AI technology from "auxiliary tool" to "scientific research partner". The large model of the Innovator scientific base is the core technology of SciMaster, and it has leading capabilities in many aspects. In terms of modal support, the model supports more than 20 scientific modalities and has general visual understanding capabilities. In terms of performance, Innovator surpasses models with 30 times the number of parameters in scientific programming tasks, demonstrating extremely high efficiency and capabilities. In February 2026, the team further released Innovator-VL, a new generation of multi-modal scientific large-scale model, providing a complete end-to-end reproducible pipeline. Among the joint publishers, Shanghai Jiao Tong University is a top university in China. It has the SJTU-DPT AI-X Lab and has profound accumulation in the field of AI for Science. Shenzhen Technology is the pioneer and leader of global AI for Science. Founded in 2018, it has completed Series C financing of more than 800 million yuan, with a post-financing valuation of more than 6 billion yuan. The company has offices and R&D centers in Beijing, Shanghai, Shenzhen, Yibin and other places. Shanghai Algorithm Innovation Institute, as a new R&D institution, provides algorithm innovation support for the project.
SciMaster has built a "read-calculate-do-write" intelligent system that covers the entire scientific research process. It is currently the only scientific research intelligent product that can cover the entire process. In terms of in-depth literature research, SciMaster can automatically search 170 million scientific research documents and global Internet resources, supports three search modes: WebSearch, WebParse, and PaperSearch, and generates in-depth research reports with literature support. All key data and core arguments have citations, which effectively suppresses the problem of large model illusion. User measured data shows that the efficiency of literature research has increased by more than 10 times. In terms of calling scientific computing tools, SciMaster supports two flexible calling mechanisms. In active calling mode, researchers actively call computing tools according to needs; in automatic calling mode, SciMaster automatically analyzes problems and intelligently matches relevant tools. This product has integrated thousands of Agent-ready tools, covering multiple subject areas. In terms of closed-loop dry and wet experiments, SciMaster can directly call the MCP service provided by Uni-Lab, seamlessly integrate laboratory instruments and software systems, and realize a complete ecosystem from software to hardware, from virtual computing to physical experiments. According to actual application cases, the cooperation with Yibin New Energy and New Materials Digital Intelligence Innovation Center has increased the efficiency of electrolyte formula R&D and delivery by 10 times. In terms of thought chain editing, SciMaster supports real-time intervention to adjust AI execution logic. Users can pause at any time during task execution, view the reasoning logic, and make real-time adjustments to task decomposition, execution sequence, key directions, etc. This unique human-machine collaboration model realizes a true "Human in the loop" interaction method. Judging from the quantitative effect, scientific research tasks that traditionally take several days to complete can be compressed to be completed within 30 minutes. Thousands of high-quality scientific research reports have been generated within one day of the product’s launch. Users with a liberal arts background have even successfully used SciMaster to complete a CRISPR gene editing review report, proving the product’s ease of use and interdisciplinary applicability.
According to current public information, the basic functions of SciMaster are currently free, and the product is in the internal testing stage. Users need to apply through the official website or obtain an invitation code to experience it. In the early days of product release, the invitation code was sold to nearly 1,000 yuan, reflecting the market's high expectations for the product. Judging from the current status, the access method is purely Web-side operation, supports mainstream browsers, does not require installation and download, and lowers the threshold for use. Users can apply for internal testing qualifications through the official website https://scimaster.bohrium.com/. Based on industry analysis and practice, SciMaster's pricing may adopt a tiered charging model in the future. Basic functions are expected to remain free, and advanced functions may adopt a subscription system or pay-as-you-go billing. The specific pricing plan has not yet been officially announced, and we need to wait for official follow-up announcements.
Judging from the user feedback collected, SciMaster has received positive reviews. Some users said that the product "makes scientific research no longer just for scientists, but also makes 'thinking like scientific research' possible for more people." In terms of application cases, the X Laboratory of Shanghai Jiao Tong University has built a "super scientific research platform" based on SciMaster. Professor Zhang Wanbin's team uses the T5-AsymCat intelligent agent to accelerate the development of chiral drug technology. At the same time, the product has reached cooperation with more than 40 universities including Peking University, Shanghai Jiao Tong University, and Xiamen University, demonstrating broad academic recognition. In terms of industry status, SciMaster has several unique advantages: its pioneering status as the world's first general-purpose scientific research agent, complete functions covering the whole process of "reading-calculating-doing-writing", relying on the professional depth of the scientific base large model Innovator, and the ecological advantage of in-depth cooperation with universities and scientific research institutions.
Compared with existing scientific research AI tools, SciMaster has significant differentiated advantages. Elicit focuses on literature review and lacks full-process coverage; ResearchRabbit focuses on literature management and lacks agent capabilities; Consensus focuses on academic search and is not a general scientific research agent; SciSpace focuses on specific scientific research stages and is not a full-process product. The uniqueness of SciMaster lies in its positioning as a truly universal scientific research agent, its native capabilities based on large scientific base models, its complete ecology of closed-loop dry and wet experiments, and its unique human-machine collaboration model.
SciMaster is mainly targeted at the following groups: university scientific researchers including graduate students, postdoctoral fellows and professors, industrial R&D engineers, interdisciplinary researchers, and scientific research managers. The product is suitable for users who need to perform full-process scientific research work such as literature research, experimental design, data analysis, and report writing. In terms of acquisition methods, users can visit the official website https://scimaster.bohrium.com/ to apply for internal testing qualifications, or obtain invitation codes through partner universities and institutions. The system supports access from mainstream browsers, no installation or download required, and easy operation.
As the world's first general-purpose scientific research AI agent, SciMaster represents an important breakthrough in the field of AI for Science. It is not only a tool, but also a changer of scientific research paradigm. By intelligentizing traditional scientific research processes such as literature research, computational analysis, and experimental design, SciMaster greatly lowers the threshold for scientific research and improves scientific research efficiency. From a technical perspective, SciMaster occupies a unique position in the field of scientific research AI based on the native capabilities of the large model of the Innovator scientific base, coupled with functional design covering the entire process. From an ecological perspective, in-depth cooperation with more than 40 universities and collaboration with Shenzhen Technology's product matrix have created strong barriers to competition. For scientific researchers, SciMaster is a tool worth paying attention to and trying, especially in scenarios that require interdisciplinary research, in-depth literature research, and complex data analysis.

User Reviews

AAgar

—

I just got the invitation code today and tried it for literature research. It’s so powerful! It automatically retrieved 170 million documents and generated reports with citation standards that were much more reliable than searching for information myself.

DebraGarciaK

—

The thinking chain editing function is really practical. You can intervene and adjust the execution logic of AI at any time to achieve true human-machine collaboration.

Carl.Collins_202050

—

Use SciMaster to write a topic review, and you can finish it in 30 minutes, which previously took several days to check. The efficiency improvement is so obvious!

RebeccaMartinezQ

—

As a liberal arts student, I was originally worried about not being able to use scientific research tools, but I easily completed the report on CRISPR gene editing, with full marks for interdisciplinary applicability.

阎婷洋

—

After being connected to the laboratory's Uni-Lab, the closed loop of dry and wet experiments is so good that the experimental equipment can be directly called to complete the verification.

Billy.LewisII816

—

Comparing Elicit and Consensus, SciMaster is the only one that covers the entire process, including literature research, computational analysis, and experimental design.

Stephen_JamesJr

—

The large-scale model based on the Innovator scientific base is indeed different. There is basically no illusion of professional terminology, and the report is highly credible.

SatsChaser453

—

It has been recommended to the students in the research group, and everyone has said it is good after using it. It is not a boast that the efficiency of literature research has increased by 10 times.

Joan_Ortiz5203

—

The research and development efficiency of electrolyte formula has been increased by 10 times. It is already being used at the Yibin Innovation Center, and the effect is really impressive.

KeithKristensen

—

It is currently free, use it and cherish it. I hope it will not be too expensive after it is officially charged.

EAdamsII

—

It has cooperated with more than 40 universities such as Peking University, Shanghai Jiao Tong University, and Xiamen University, and is highly recognized by the academic circle.

Sophia_Ward_99

—

Supports three search modes (WebSearch, WebParse, PaperSearch), which can be flexibly selected according to needs.

Margaret.Scott

—

All key data in the generated report have citations, so you no longer have to worry about the illusion of a large model.

星辰_12

—

Integrating thousands of Agent-ready tools, the scientific research tool ecosystem is very complete.

EtherDream366

—

Shenzhen Technology has just completed a Series C financing of 800 million, with a valuation of over 6 billion, and it has strong support behind it.

Jesse.WardQ

—

It supports more than 20 scientific modes, has strong general visual understanding capabilities, and comprehensive coverage of scientific research scenarios.

LoganRodriguez_20234

—

There is no need to install or download, it can be used directly by accessing the browser, which is very friendly to Mac users.

wu1oi

—

Professor Zhang Wanbin's team is already using it. Accelerating the development of chiral drug technology is indeed a change in the scientific research paradigm.

骑士602

—

The vision of making scientific research no longer just for scientists and allowing more people to think and work like scientists touches me very much.

MateoNielsen

—

The world's first general-purpose scientific research agent is indeed an important breakthrough in the field of AI for Science and a milestone product.

SciMaster

In-depth Report

User Reviews

Similar Products