Want a smart insight into your inbox? Sign up for our weekly newsletters to get the only thing that is important to enterprise AI, data, and security leaders. Subscribe now
Chinese AI Startup FamiliarMulti-agent for consumers, who have made headlines earlier this year for their approach to the multi-agent orchestration platform and “pro”-soomers (wishing to operate work works), have returned with an interesting new use of their technology.
While many other major rivals AI providers such as Openi, Google, and Ze It has launched a “deep research” or “deep researcher” AI agents who conduct a wide, deep web research of minutes or hours and present full reports from consumers, Manas are taking a different approach.
The company just announced a “vast research”, A new experimental feature that enables consumers to take advantage of the strength of parallel AI agents, enables users to implement large, high volume tasks-more than 100, at the same time, are focused on completing the same task (or a series of sub-works stairs is maximum purpose).
Earlier, Menus was told that he was using the Anthropic Claude and Alibaba Kevin model to strengthen his platform.
AI Impact Series returning to San Francisco – August 5
The next step of the AI is here – are you ready? Block, GSK, and SAP leaders include for a special look on how autonomous agents are changing enterprise workflows-from real time decision-making to end to automation.
Now secure your place – space is limited:
Parallel processing for research, abstract and creative output
A Posted on Video Official X accountMans co -founder and chief scientist Yacho ‘Peak’ shows the demo of extensive research use to compare G100 shoes.
To complete this task, Menus Wide Research almost immediately rotates 100 harmonious subjects – each shoe is assigned to analyze the design, pricing and availability of each shoe.
The result is a setting matrix that is provided in minutes in both the spreadsheet and webpage forms.
The company suggests that extensive research is not limited to just data analysis. It can also be used to find creative tasks such as design.
In a scenario, Menus agents developed posters designs in 50 separate visual styles simultaneously, returning the polished assets to the downloadable zip file.
According to the Menus, this flexibility is from the level of the system level to the parallel processing and agent to the agent communication.
In the video, Peak explains that extensive research is a better virtualization and agent architecture, a better virtual and agent architecture capable of scaling computing power 100 times from preliminary offerings.
This feature is designed to automatically activate during tasks that require widespread analysis, without any manual toggle or formation.
Availability and pricing
Extensive research is available for users today on the Menos Pro Plan and will gradually become accessible to plus and basic projects. So far, subscription pricing for Mans is following every month.
- Free -/0/month includes 300 daily refresh credit, chat mode access, 1 seat work, and 1 scheduled task.
- Basic – $ 19/monthly 1,900 monthly credit (+1,900 bonuses), 2 harmony and 2 scheduled work, agent mode, image/video/slides of the Slides Generation, and special data sources have been added.
- Plus – $ 39/month 3 sesame and 3 scheduled work, 3,900 monthly credit (+3,900 bonuses), and includes all the basic features.
- De -$ 199/month offers a complete feature set, including 10 harmony and 10 scheduled task, 19,900 credit (+19,900 bonus), preliminary access to beta features, a manus T -shirt, and the production of the latest agent tools and content production.
There is also a 17 % discount on the prices that are for the consumers who want to pay the annual front.
The launch is based on the infrastructure introduced with Manas earlier this year, which the company describes as not only as an AI agent, but also a personal cloud computing platform.
Each Monos session runs on a dedicated virtual machine, which gives users access to archetypel cloud computers through natural language-a setup views the company as a key to activating the real general purpose AI workflow.
With extensive research, Menus users can hand over dozens of or even hundreds of subjects to research or creative search.
Unlike the traditional multi-agent system with a default role (such as manager, coder, or designer), each subject within extensive research is a fully qualified, fully featured Menus example-not a special character not a special-work-free-working and performing any common work.
The company says the architectural decision opens the door to handle flexible, expanding work, which is unorganized by strict templates.
What are the benefits of deep research?
This means that running all these agents parallel is faster, and that will result in a better and more diverse set of work products above the research reports, as shown by other AI providers or landed.
But although the manus agent promotes extensive research as a breakthrough in harmony, the company does not provide direct evidence that spreading dozens or hundreds of subjects is far more efficient than handling the same, high -capable agent tasks.
The release does not include performance standards, comparisons, or technical explanations to justify the business relationship of this approach-such as increasing the use of resources, coordination complexity, or potential incompetence. It also lacks the details of how the Subaginants cooperate, how the results are integrated, or this system offers a measured benefits to speed, accuracy or price.
As a result, while the feature exhibits architectural ambitions, but its practical benefits are in easy ways based on the information provided.
Sub -agents usually record a mixed track, so far …
Although the implementation of a wide range of measures is as progress in the general AI agent system, the broader ecosystem has seen mixed results with similar subjective perspectives.
For example, on Self -described users of Reddate, Claude Code It has raised concerns about the slowdown of the subsidiaries, the use of large amounts of token and the offering of limited exposure to its implementation.
Ordinary pain points include abnormal performance during coordination protocol deficiency, debugging problems, and high load periods.
These challenges do not necessarily consider the implementation of the Menus, but they highlight the complexity of developing a strong multi -agent framework.
Menus acknowledges that extensive research is still experimental and can come up with some limitations in relation to development.
Are looking forward to
With a wide research rollout, Menus deepened his commitment to explain how consumers interact with AI agents on a scale.
Since other platforms are wrestling with the technical challenges of subsidiary and reliability, the vision of the man can act as a test case whether ordinary agents can easily provide scope modules instead of scope modules-smooth, multilateral AI cooperation.
The company pointed to broader ambitions, which show that the infrastructure behind extensive research is the basis of future offerings. Consumers and industry watchdog will equally focus on whether this new wave of agent architecture can remain in accordance with its abilities – or whether the challenges shown somewhere in the AIA will eventually be overcome.
