Zero and One, all things
As a pioneer in AI 2.0 technology, Zero One Ten Thousand is committed to driving innovation and the application of artificial intelligence in order to bring its benefits to everyone. By offering commercial solutions, sharing technical knowledge, bringing together talented professionals, and fostering open cooperation, the company is helping to shape AI 2.0.
Tags:AI training modelsWhat is Zero One Ten Thousand Things?
Zero One Everything 01.AI is an artificial intelligence company that provides large models, agents, and AI transformation services for corporate and industrial applications.
The focus of the current official website has shifted from individual online modeling experiences to AI for business decision-making, industry-specific solutions, and collaborative delivery.
Current products and services
- Wan Ce Enterprise’s AI decision-making platform.
- AI designed for the top executives of companies.
- AI for investment officers in investment and strategic contexts.
- Top-selling AI for B2B and B2G sales.
- A one-stop platform for MindWorks’ large-scale models.
- Strategic consulting services for AI transformation.
- Frontier Deployment Engineers (FDE) collaborate to ensure successful delivery.
- Sovereign AI and industry-specific solutions.
- The Yi series of open-source large language models.
WanCe Decision-Making Hub Platform
WanCe is designed for high-value decision-making scenarios in areas such as corporate strategy, management, sales, and investment, emphasizing the transition from information analysis to actionable steps.
| Ability | Main function | Enterprise value |
|---|---|---|
| Integration of operational facts | Connect internal and external information within the enterprise | Reduce information dispersion |
| Risk identification | Identifying anomalies, assumptions, and decision-making blind spots | Exposing problems in advance |
| Chain of evidence | Retain the basis for analysis and citation relationships. | Facilitates review and judgment |
| Key questions | The conclusions and gaps in the challenging materials | Improve the quality of decisions |
| Action closure loop | Convert the conclusions into tasks and keep tracking them. | Drive the implementation of decisions |
Decision-making AI is suitable for helping managers identify problems and gather evidence; it should not replace statutory approvals, professional judgment, or the person who bears ultimate responsibility.
Boss AI
Boss AI is designed for senior corporate executives; it organizes corporate information around business analysis, risk identification, and action implementation.
- Summarize the key operational facts.
- Identify the main risks affecting the target.
- Compare different options and resource allocations.
- Formulate actionable recommendations.
- Track tasks, responsible persons, and outcomes.
- Retain the decision-making process for managing post-mortems.
Investment Officer AI
The AI service for investment officers supports investment decisions, strategic planning, and board decisions, with a focus on material analysis, risk assessment, and addressing key challenges.
| Scene | AI-assisted tasks | Manual confirmation is required. |
|---|---|---|
| Industry research | Organize market, company, and competitive information | Data timing and source |
| Initial project screening | Extract the business model and key metrics | Authenticity of materials |
| Risk review | Identify hypothesis conflicts and information gaps. | Legal, financial, and technical due diligence |
| The selection committee is getting ready. | Create a summary and a list of follow-up questions | Final investment decision |
| Post-investment tracking | Monitoring indicators and operational changes | Causes of abnormalities and corrective actions |
The investment officer AI is not a licensed investment advisor, and the conclusions derived from its models cannot serve as the sole basis for making securities transactions or major investment decisions.
Xiaoguan AI
PinGuo AI is designed for complex B2B and B2G sales management, helping teams assess the quality of business opportunities and identify the appropriate solutions.
- Analyze customer needs and the decision-making process.
- Assess the deal stage and the probability of closing the deal.
- Identify budget, competition, and implementation risks.
- Match product capabilities with customer issues.
- Optimize the priority of investment in pre-sales resources.
- Summarize experiences gained and improve sales strategies.
WanZhi Enterprise Large Model Platform
WanZhi is a one-stop platform for enterprise-level large models; its official website describes it as tools for model deployment, practical application, and model fine-tuning.
| Hierarchy | Possible tasks to be covered | Confirm at the time of purchase |
|---|---|---|
| Model deployment | Connect to or deploy the selected model | Cloud, private cloud, or on-premises settings |
| Application practice | Build knowledge bases, agents, and business applications | Current features and customizable boundaries |
| Model fine-tuning | Use enterprise data adaptation tasks. | Volume of data, computing power, and evaluation methods |
| Security governance | Permissions, auditing, and data isolation | Compliance proof and log scope |
| Operational assessment | Monitor quality, cost, and business metrics | SLA and the responsibility for continuous iteration |
Strategic consulting and co-creation with FDE
Zero One Everything employs a approach that combines strategic consulting with FDE co-creation in order to link business objectives, data, models, and application deployment.
- The top executive of the company defines the business objectives and boundaries.
- Identify high-value use cases, risks, and expected ROI.
- Design a roadmap, technology stack, and delivery milestones.
- Build a business ontology for corporate entities, processes, and events.
- Access data and handle permission and quality issues.
- The prototype is developed jointly by FDE and the business team.
- Verify accuracy, efficiency, and cost using real tasks.
- Establish mechanisms for manual review, auditing, and exception handling.
- Monitor business metrics continuously after going live.
- Use scenario experience to feed back into model development and strategic iteration.
Industry solutions
| Industry or field | Typical problems | Entry points for AI |
|---|---|---|
| Sovereign AI | National-level autonomy in data and capabilities | Secure base and local capabilities |
| Supply chain logistics | Complex scheduling and changes in supply and demand | Prediction, optimization, and decision support |
| Manufacturing | Coordination of production, quality, and equipment | On-site perception and operational decision-making |
| Energy | Supply-demand balance and scheduling | Prediction and optimization suggestions |
| Agriculture | Data at the production site is scattered. | Crop and livestock management as well as intelligent operations |
| Investment | A large number of materials and complex risks | Evaluation, review, and follow-up questions |
| Education | Personalization in teaching and management | Capabilities of teaching assistants, teachers, and mentors |
| Retail | Coordination of people, products, and venues | Operational analysis and strategy optimization |
Notice on the suspension of services for the open platform
The Zero-One Everything Large Model Open Platform has announced that it will gradually cease to offer users online access, API calls, and top-up services.
- The old API prices should no longer be listed as the current valid prices.
- New projects should not rely on the payment methods that are about to stop being available.
- Existing users should export the necessary data and bills as soon as possible.
- Check the existing API keys, balance, and service duration.
- Evaluate migrating to other models or self-hosted Yi models.
- For special corporate needs, please contact the official business team.
Yi series open-source models
The official GitHub page of ZeroOneTen Thousand has made available the code for the Yi and Yi-1.5 series of models, information regarding their weights, as well as materials on fine-tuning, quantization, and deployment.
| Model series | Main positioning | Typical context |
|---|---|---|
| Yi-6B | Lighter base and dialogue tasks | 4K and extended versions |
| Yi-9B | Code, mathematics, and general reasoning | Includes 200K version |
| Yi-34B | Foundation and dialogue models with higher capacity | Includes 200K version |
| Yi-1.5 | Improves coding, mathematics, reasoning, and instruction-following skills. | Select by specific model card |
| Yi-VL | Image and Text Understanding | By specific version |
| Quantitative model | Reduce the requirements for video memory and deployment. | AWQ or GPTQ version |
These are mainly models released between 2023 and 2024; when selecting one, it is necessary to compare its quality, context, ecosystem, and maintenance status with those of the newer models available today.
Open-source license
The official repository states that the code and weights related to the Yi-1.5 series are licensed under the Apache 2.0 license, and can be used for personal, academic, and commercial purposes.
- Commercial use still requires compliance with the full license text.
- Derivative works should retain the attribution requirements specified by the authorities.
- Different warehouses, weights, and data may have separate terms.
- Third-party quantitative models do not automatically inherit the official credibility.
- Before going live, check the NOTICE, model cards, and dependency licenses.
- The fact that it is an open-source model does not mean that the officially hosted API is still available for purchase.
Local deployment method
The official repository provides reference paths for Transformers, Docker, llama.cpp, quantization, fine-tuning, and web demos.
Yi Open-Source Model Deployment Tutorial
- First, choose between a basic, chat, or visual model.
- Read the corresponding model card and license.
- Check the requirements for video memory, RAM, CUDA, and disk space.
- Obtain weights from official repositories or model platforms.
- Deploy the inference in an isolated testing environment.
- A evaluation set is created using real business samples.
- Test hallucination, security, latency, and concurrency.
- Use AWQ or GPTQ for quantization when necessary.
- Check the rights and quality of the training data before fine-tuning.
- After going live, continuously monitor outputs and dependent vulnerabilities.
Refer to the hardware requirements.
The official older documentation provides a minimum requirement for video memory; the actual amount needed depends on factors such as precision, context length, concurrency, and the inference framework used.
| Model | Minimum VRAM requirement per official documentation | Deployment tips |
|---|---|---|
| Yi-6B | About 15GB | Quantification can further reduce it. |
| Yi-9B | About 20GB | A 24GB graphics card makes testing easier. |
| Yi-34B | About 72GB | Multiple GPUs or GPUs with large memory capacities are usually required. |
| Yi-6B-200K | About 50GB | Long contexts significantly increase resources. |
| Yi-34B-200K | About 200GB | A high-end multi-card environment is required. |
The above information is based on historical data from the official warehouse; it does not represent fixed CVM configurations or current purchase prices.
Price and procurement methods
The official website does not disclose the standard pricing for the company’s products; typically, quotes are required based on the specific scenario, data volume, deployment needs, and scope of services.
| Usage method | Price of software or services | Other costs |
|---|---|---|
| Yi open-source model | Weight download is free. | Computing power, storage, operation and maintenance, and security |
| Open platform API | Recharges and calls are being gradually discontinued. | It is not recommended to plan new projects based on the old prices. |
| Wan Ce and Character AI | Contact the authorities to inquire about prices. | Integration, data governance, and delivery |
| WanZhi Enterprise Platform | Contact the authorities to inquire about prices. | Deployment, tuning, and operation |
| Consulting and co-creation with FDE | Quotation by project | Scope, cycle, and on-site resources |
When purchasing services, businesses should request a list of functions, delivery boundaries, SLAs, data-related terms, acceptance criteria, and plans for exiting the migration process.
Corporate procurement checklist
- Identify the business metrics and baselines that need to be improved.
- Determine whether the product is a standardized service or a custom project.
- Verify the cloud, private cloud, and on-premises deployment options.
- Confirm the data location, training usage, and deletion mechanisms.
- Define the model version as well as the plans for upgrading and rolling back.
- Testing permissions, auditing, content security, and manual review.
- Define fault response, availability, and compensation boundaries.
- Calculate the total cost of computing power, implementation, operation and maintenance, as well as subsequent iterations.
- Develop a plan for the exit of the design supplier and for data migration.
Which users are it suitable for
| User | Recommended directions | Unsuitable expectations |
|---|---|---|
| Corporate management | Boss AI and Wan Ce Decision-Making Center | Let AI assume decision-making responsibilities on its own. |
| Investment firms | Investment Officer AI and Material Analysis | Replacing due diligence and the investment committee |
| Complex sales team | Top-performing AI and business opportunity management | Relying solely on the model to ensure a deal is closed |
| Technical teams in large enterprises | Co-creation by Magic Platform and FDE | Go live with zero integration required |
| Developers and researchers | Local deployment of Yi open-source model | Continue to rely on the old open platform for top-ups. |
Usage restrictions and risks
- There are no fixed, public prices for corporate services.
- The original open platform is stopping its online experience, API services, and top-up options.
- The main models released in the open-source repository are not the latest generation.
- Large models may generate incorrect and unreasonable content in their outputs.
- A context of 200K does not mean that all locations are equally reliable.
- High-risk decisions must be reviewed by professionals.
- Local deployment requires computing power, operational capabilities, and security measures.
- The effectiveness of implementing industry solutions depends on data and process transformations.
Frequently Asked Questions
What services does Zero One Ten Thousand offer primarily at the moment?
The current focus areas are WanCe’s enterprise decision-making platform, AI for business owners, AI for investment officers, AI for top sales performers, WanZhi’s enterprise large-model platform, as well as consulting services and collaborative development via FDE.
Can the Zero One All Things Open Platform still be used for top-ups and API calls?
According to the official announcement, the platform will gradually cease to offer online trials, API calls, and top-up services; new projects should not be developed following the old service models.
Is the Yi large model open-source?
Yes, the official GitHub repository makes models such as Yi and Yi-1.5 available; the code and weight files for Yi-1.5 are licensed under the Apache 2.0 license.
Can the Yi open-source model be used for commercial purposes?
The official repository states that it can be used for commercial purposes, but it is necessary to comply with the licensing requirements, include the required attribution information, and check the specific weight constraints and dependency rules.
How much are the products of Zero One Ten Thousand Enterprises?
The official website does not disclose fixed pricing for WanCe, WanZhi, and character AI; it is necessary to contact the authorities to request a quote based on the deployment, data volume, integration requirements, and scope of delivery.
Can the Yi model be deployed locally?
Yes, the official repository provides Transformers, Docker, llama.cpp, as well as materials related to quantization and fine-tuning; however, it is necessary to have sufficient video memory and the necessary operational capabilities.
Is Zero One Everything suitable for individual users?
Current commercial websites are primarily aimed at businesses; individual developers would be better off studying the Yi open-source models, rather than relying on open platforms that are no longer in use.
A tool similar to something that represents zero and ten thousand things.
Guigong Network Security Registration No. 45132202000164