What is Qianwen Cloud
Qianwen Cloud is Alibaba Cloud's all-new platform built for AI AgentsMaaS model service platform, aggregating over 150 mainstream large model APIs, and supporting the comprehensive encapsulation of model services into Skills and CLI tools, enabling Agents and developers to complete model selection, invocation, and usage management with a single command, driving AI productivity into the Agentic era.
Qianwen Cloud's main features
- One-stop access via the Model Plaza: Aggregates over 150 mainstream and third-party models including Qwen, GLM, Kimi, DeepSeek, Wan, HappyHorse, and more, covering all modalities of text, images, video, and speech.
- Multi-dimensional model comparison and selection: Supports side-by-side comparison by attributes such as parameter scale, capability dimensions, price, context length, modality support, and applicable scenarios; after comparison, you can go directly to the experience page to verify output results with real prompts.
- Skills skill package: Encapsulates the complete chain of model selection, authentication configuration, API invocation, and usage queries into skill combinations that Agents can understand. Agents can dynamically route different models and automatically invoke capabilities such as vision, image generation, and video, with no need for users to write code.
- CLI command-line tool: Provides command-line capabilities such as login authentication, model queries, model invocation, usage queries, and environment diagnostics, and supports automating all model service workflows through scripts.
- Intelligent usage management: Agents can pull model usage data in real time, automatically analyze consumption trends, identify anomalies, and output cost optimization suggestions, with process data transparently presented through CLI logs.
- Token Plan subscription: Introduces subscription-based plans for high-frequency AI programming and Agent tool scenarios, with one API Key providing all-modality access, compatibility with mainstream AI tools, controllable monthly fees, and costs significantly lower than pay-as-you-go.
How to use Qianwen Cloud
- Visit the official website: Open the Qianwen Cloud official website https://www.qianwenai.com/ , browse the Model Plaza, and compare the parameters, capabilities, and prices of the models you need.
- Install the CLI tool: In the terminal, run
npm install -g @qianwenai/qianwen-clito install the global command-line tool. - Configure Skills: Run
npx skillsadd QianWen-AI/qianwen-aito install the Qianwen Cloud skill package for the Agent, or visit the Skills.md documentation provided by the platform and configure it according to the instructions. - Authentication and invocation: Complete login authentication through the CLI. After obtaining the API Key, you can use a single command to let the Agent automatically complete model selection, invocation, and task execution.
- Usage monitoring and optimization: View model usage, bills, and free quotas in real time through the CLI, and the Agent will automatically analyze trends and provide cost optimization suggestions.
Qianwen Cloud's core advantages
- China's first model service platform deeply oriented toward Agents: Shifts from "serving people" to "serving Agents," reconstructing the core interaction paradigm of the model platform.
- Open-source ecosystem: The Skills and CLI toolkits are already open source on GitHub, and developers can freely extend and integrate them.
- Multi-model compatibility: One API Key provides access to over 150 models, avoiding switching between multiple platforms and managing multiple keys.
- Extreme cost-effectiveness: Token Plan subscriptions are as low as 40% off, with input for a million Tokens starting at only 1 yuan, significantly reducing costs for high-frequency Agent scenarios.
- Flexible human-machine collaboration: Supports autonomous invocation by Agents, while retaining interactive entry points for manual online trials, comparison and selection, and Prompt debugging.
Comparison with similar competing products of Qianwen Cloud
| Dimension | Qianwen Cloud (Alibaba Cloud) | iFlytek Spark MaaS(iFlytek) |
|---|---|---|
| Product positioning | A MaaS service platform born for AI Agents, comprehensively refactored for Agents from atomic capabilities to interaction logic | A one-stop large model customization and fine-tuning platform, focusing on full-chain engineering of data management, model fine-tuning, evaluation, and deployment |
| Model ecosystem | Aggregates more than 150 mainstream and third-party models, covering full modalities such as Qwen, GLM, Kimi, DeepSeek, Wan, and HappyHorse | Integrates more than 50 models, with the iFlytek Spark series of self-developed models as the core, and is compatible with open-source models such as Llama 3, Qwen, and DeepSeek |
| Agent-native support | Native Skills skill packages and CLI tools; Agents can learn the entire site's capabilities with a single instruction, enabling dynamic model routing and autonomous invocation | Provides OpenAI-compatible interfaces, supports integration with programming tools such as OpenClaw, Cursor, and Claude Code, and launches Coding Plan targeted services for Agent programming scenarios |
| Interaction method | Three-end collaboration of web-based multi-dimensional comparison experience + Skills skill packages + CLI command line, allowing both humans and Agents to invoke flexibly | Mainly Web visual console + API invocation + Coding Plan packages, focusing on manual configuration and code tool integration |
| Model customization capability | Centered on model selection, invocation, and management, it does not provide native fine-tuning features and focuses on efficient model use by Agents | Its core advantages are zero-code model fine-tuning and data engineering, supporting LoRA, full fine-tuning, data cleaning and enhancement, Loss curve monitoring, and one-click deployment |
Application scenarios of Qianwen Cloud
- AI programming and development: Developers useQoderand other programming Agents to directly call Qianwen Cloud CLI, automatically completing model selection, code generation, and deployment.
- Multimodal content creation: Leveraging Wanxiang capabilities, Agents automatically schedule text-to-image, text-to-video, and image-to-video models to achieve film-grade batch content production.
- Enterprise intelligent office: IntegratesWukongand other work assistants; with one sentence, multiple applications can be mobilized to complete tasks such as scheduling, documents, approvals, and meetings.
- Intelligent cockpit and travel: Leveraging the Qianwen large model to integrate capabilities such as travel assistants, intelligent planning, and long-term memory, building a safe and pleasant in-vehicle interaction experience.
- Social and virtual companionship: For social anthropomorphic interaction scenarios, it supports virtual IP creation and real-time emotional dialogue, providing a new immersive interaction experience.
Frequently asked questions about Qianwen Cloud
Q: Can the same subscription be used in multiple tools?
- Yes. The same API Key can be used in all compatible AI tools, with shared quota consumption.
Q: Does Token Plan support cancellation?
- The Token Plan team edition does not support cancellation.
Q: Can the API Key of the Token Plan team edition be mixed with other API Keys?
- No. The API Keys and Base URLs of the Token Plan team edition, Coding Plan, and pay-as-you-go billing are not interconnected. Do not mix them. Misusing another API Key will not deduct the package quota of the Token Plan team edition.
Q: What are the usage restrictions of the Token Plan?
- It is limited to interactive use in compatible AI programming and agent tools, and may not be used for automated scripts or application backends. Violating these terms may result in subscription suspension or API Key ban.
