DEV Community: liekeai

Agent Mail: Tencent Just Gave AI Agents Their Own Mailbox

liekeai — Sun, 18 Oct 2026 04:47:03 +0000

Agent Mail: Tencent Gives AI Agents Their Own Mailbox (@agent.qq.com), Open Source & Free in Beta | Lieke AI

@agent.qq.comDedicated agent mailbox Apache-2.0Open source 50/dayBeta send quota 6+ platformsMajor agents supported

In one line: why AI agents need their own mailbox

Two old paths for letting agents handle email — grant access to your personal inbox (privacy risk) or run your own SMTP (ops cost). Tencent QQ Mail's answer: give the agent a digital identity independent from humans: an isolated xxx@agent.qq.com address that can never touch your personal QQ/WeChat mail.
Internal beta opened June 23, 2026. Source code is Apache-2.0 on GitHub (Tencent/AgentlyMail); official site: agent.qq.com.

3-step setup, tested in ~5 minutes

Let the agent install it: ask your agent to read https://agent.qq.com/doc/cli-setup.md and follow the steps to install & configure Agent Mail CLI.

It runs automatically:npm install -g @tencent-qqmail/agently-clinpx skills add Tencent/AgentlyMail -g -yagently-cli auth loginOpen the printed authorization link in a browser.
Scan to authorize: after OAuth, run agently-cli +me — it returns xxx@agent.qq.com when configured.

Works with WorkBuddy, Claude Code, Cursor, Kimi Work, Doubao, Codex, Hermes and other mainstream agents. Read, send, search, attachments and live watch — all via one CLI.

Security: three layers built for agent workloads

Layer	What it does
Physical isolation	Agent gets a fresh data space; even a compromised agent cannot reach personal mail
Two-stage confirmation	Send/reply/forward/delete requires the agent to generate a summary + confirmation token, then the user approves
Prompt-injection defense	Malicious instructions inside inbound mail are treated as plain text, never as system commands

Six practical use cases (hands-on)

Invoice auto-expense: agent extracts amount/tax ID and files reimbursement reports
Newsletter digest: forward newsletters, get a daily AI summary
Order reconciliation: agent pulls order numbers into your sheets
A2A business email: procurement agent ↔ supplier agent, fully auditable
Verification-code relay: use Agent Mail for signups, agent auto-fills codes
AI job applications: dedicated address for resumes and interview scheduling

Competitive landscape: the agent-mail race

	Agently Mail (Tencent)	AgentMail (YC)	Mail4Agent
Integration	WeChat scan + CLI	API + SDK	API-first
Pricing	Free in beta	3 mailboxes free, $20-200/mo paid	Not public
Security	Two-stage confirm + injection defense (unique)	DIY	DIY
China access	No VPN needed	VPN needed	Fair
Ecosystem	CLI + Skill path	Python/TS SDK, MCP Server, Webhook	Multi-framework

Bottom line: Agently Mail is the lowest-friction option for CN developers; AgentMail suits large overseas API/SDK deployments.

Beta limitations (stated honestly): 50 sends/day, 1GB per address — not production-ready until quotas open up; official Skill has Windows HTML-email layout issues (community fixes exist); A2A cross-domain protocol not yet documented; mailbox prefix can be changed only once.

Before your agent can run 24/7, you need reliable cloud compute

Agently Mail is just the first piece of agent infrastructure. The real backbone is elastic cloud servers that keep agents online around the clock. Lieke AI helps you get official promotional pricing.

Check Alibaba Cloud ECS deals →

FAQ

Does authorization leak my QQ password? No. It uses the standard OAuth device-code flow; the CLI never stores your QQ password.
Install fails or is slow? Make sure npm can reach the public registry; if the auth link won't open, check access to agent.qq.com.
How does this differ from AgentMail / Mail4Agent? Same category, different positioning: Agently Mail focuses on China accessibility + CLI/Skill + unique security; AgentMail targets overseas API/SDK scale.
Production-ready? Beta quotas are tight (50/day, 1GB). Start with daily reports and newsletter digests; wait for quota release for production scale.

Sources: agent.qq.com · GitHub · CSDN hands-on review · CSDN CLI guide · GMT EIGHT

Hosting Production APIs on Cloud Servers: Architecture and Best Practices

liekeai — Tue, 29 Sep 2026 04:47:03 +0000

Hosting Production APIs on Cloud Servers: Architecture and Best Practices

APIs have different operational demands than websites. Learn the patterns that keep production APIs fast, available, and safe under real traffic.

Design for statelessness

Stateless API servers scale horizontally and fail gracefully: any instance can serve any request. Keep sessions and cache in shared services such as Redis, and uploads on object storage.

Make deployments blue-green or rolling so you can ship without dropping in-flight requests. Health checks should verify dependencies, not just process liveness.

Rate limiting and abuse protection

Put rate limiting at the gateway or load balancer before it reaches your application. Set limits per key or per IP, and return proper 429 responses with Retry-After headers.

Add basic abuse signals: block known bad actors, validate input strictly, and monitor for unusual traffic patterns that could signal scraping or attacks.

Reliability patterns

Use timeouts and circuit breakers for all downstream calls. A slow dependency should degrade gracefully, not pile up requests and take the whole service down.

Design idempotency for writes so retries are safe. Queue background work where possible so bursts do not force you to scale for the peak.

Observability

Log structured requests (method, path, status, latency, trace ID) and ship them to a central store. Expose metrics for traffic, errors, latency and saturation — the four golden signals.

Alert on error rate and p95 latency, not just CPU. Error-rate alerts catch incidents that load-based alerts miss entirely.

Frequently Asked Questions

What size server do I need for an API?

Start small — most APIs run fine on 1-2 vCPU with 2-4GB RAM. Measure latency and error rate under load, then scale based on data, not intuition. Keep the database separate.

How do I handle traffic spikes?

Use auto-scaling based on CPU or request queue depth, keep the database scaled for the peak or use read replicas, and offload bursts with a queue where possible. Test with load testing before real spikes arrive.

Should my API be behind a gateway?

Yes, for production: a gateway or load balancer gives you TLS termination, rate limiting, routing, and a single entry point for observability. It also lets the application layer focus on business logic.

Prices and promotions shown on this page are indicative only — always confirm current pricing and terms on the official Alibaba Cloud website at the time of purchase.

More cloud deals and independent dev-tool guides: lieke-ai.com — Alibaba Cloud international coupons · China new-user deals.

Alibaba Cloud ECS vs AWS EC2: A Migration-First Comparison for 2026

liekeai — Mon, 28 Sep 2026 04:47:01 +0000

Alibaba Cloud ECS vs AWS EC2: A Migration-First Comparison for 2026

Both platforms are excellent, but they are optimized for different teams. This guide compares the practical differences — regions, billing models, ecosystem and migration effort — so you can choose the right home for your workloads.

Why the choice matters more than raw specs

Cloud servers look similar on paper: CPU, memory, storage, bandwidth. In practice, the differences that affect you are regional availability, networking quality from your users, the billing granularity, and how much operational tooling you get out of the box.

Alibaba Cloud is the leading cloud in Asia-Pacific and has strong presence in Europe and North America, while AWS offers the broadest global footprint. If most of your users are in Asia, Alibaba Cloud regions are often closer and the network quality to China — for cross-border applications — is a real advantage.

Billing and cost structure

AWS charges by the hour with a long list of instance families and per-second billing on many services. Alibaba Cloud offers both pay-as-you-go and subscription models, and for stable workloads the subscription discounts are generally deep.

A practical approach: estimate your steady-state load, compare the subscription price after discounts, and keep pay-as-you-go for burst or test environments. Do not compare list prices only — new-user offers and long-term commitments change the picture dramatically.

Ecosystem and tooling

Both clouds have mature ecosystems: load balancers, managed databases, container services, serverless, monitoring, and AI services. Alibaba Cloud’s offering is very complete for web, e-commerce, gaming and AI workloads, and its pricing for managed database and storage services is often competitive.

For teams already using open-source tooling (Terraform, Kubernetes, Prometheus), both clouds work fine. For teams that want a single console experience, both are equally valid — the deciding factor is usually regional performance and total cost.

Migration effort and risk

Migrating between clouds is mostly about data transfer, re-pointing DNS, and re-testing integrations. Keep a parallel run during migration, use database replication instead of cold copy, and validate latency from your real user locations before cutting over.

A migration is also a good moment to clean up: right-size instances, consolidate storage classes, and add monitoring from day one.

Frequently Asked Questions

Is it easy to move from AWS to Alibaba Cloud?

Yes. Both support standard Linux images, Docker, Kubernetes and Terraform. Plan a parallel run, replicate your database, point DNS after validation, and keep the old environment as a rollback target for at least two weeks.

Which is cheaper for a startup?

It depends on your region and workload type. For Asia-focused startups, Alibaba Cloud’s subscription plans and new-user offers are usually very competitive. Always compute the total cost including bandwidth and managed services, not just the instance price.

Can I run the same architecture on both?

In almost all cases, yes. Standard components such as Nginx, PostgreSQL, Redis and Kubernetes are provider-agnostic. The main work is adapting provider-specific services such as DNS, object storage SDKs and IAM policies.

Prices and promotions shown on this page are indicative only — always confirm current pricing and terms on the official Alibaba Cloud website at the time of purchase.

More cloud deals and independent dev-tool guides: lieke-ai.com — Alibaba Cloud international coupons · China new-user deals.

Building an AI Inference Server: GPU Choices and Cloud Deployment

liekeai — Sat, 26 Sep 2026 04:47:03 +0000

Building an AI Inference Server: GPU Choices and Cloud Deployment

Deploying AI models for real users requires thinking about GPU capacity, latency, cost and scaling. Here is a practical guide to running inference servers on the cloud.

GPU selection and right-sizing

Start by benchmarking your model, not by picking the biggest GPU. Measure tokens per second, batch throughput, and memory footprint for your actual workload, then choose the smallest GPU class that meets your latency target.

For LLM serving, memory capacity and bandwidth matter more than raw compute. For image or video models, compute throughput dominates. Different model families justify different instance types.

Serving architecture

Use a dedicated inference server (vLLM or similar) rather than calling the model ad hoc from application code: request queueing, continuous batching, and P50/P95 latency control come free.

Put the inference tier behind a queue or load balancer with concurrency limits so a traffic spike degrades latency gracefully instead of OOM-ing the GPU.

Cost control for GPU workloads

GPUs are the most expensive cloud resource you will rent. Use spot or discounted capacity for batch and non-critical inference, and reserve committed instances for steady production load.

Scale to zero when idle for dev and staging environments. For spiky production traffic, evaluate pay-per-token API services versus self-hosted GPUs — the crossover point depends on your volume and model size.

Operational considerations

GPU instances need more careful health monitoring: temperature, VRAM usage, and error counters. Set alerts before they become outages.

Keep model weights in object storage and cache them on the instance disk to avoid slow cold starts. Version models and track which version is serving so rollbacks are one command.

Frequently Asked Questions

How much GPU memory do I need?

For open-weight LLMs, a common rule is roughly 2x the model size in GB of VRAM for fp16 weights plus overhead for KV cache. Benchmark with your real model and context length before committing.

Self-hosted GPUs or API services?

For low or spiky volume, API services win on cost and effort. For sustained high volume or strict data requirements, self-hosting wins. Calculate the crossover at your actual usage, including engineering time.

How do I reduce inference latency?

Use batching for throughput, keep the model warm, place the region close to users, and consider model quantization or distillation if latency remains too high after these steps.

Prices and promotions shown on this page are indicative only — always confirm current pricing and terms on the official Alibaba Cloud website at the time of purchase.

More cloud deals and independent dev-tool guides: lieke-ai.com — Alibaba Cloud international coupons · China new-user deals.

Alibaba Cloud Wuying Cloud Desktop Review 2026: Performance, TCO & Real-World Use Cases

liekeai — Fri, 25 Sep 2026 04:47:02 +0000

Alibaba Cloud's Wuying (Shadowless) Cloud Desktop has evolved from a niche VDI solution to a mainstream cloud computing platform. In 2026, with AI acceleration features, GPU options, and an expanding ecosystem, Wuying is redefining what "cloud desktop" means. This in-depth review covers performance benchmarks, total cost of ownership compared to traditional PCs, and which businesses benefit most from making the switch.

Try Wuying Cloud Desktop — Free Trial Available

Experience enterprise-grade cloud desktops with exclusive channel pricing

Start Free Trial →

Explore Wuying Plans →

What Is Wuying Cloud Desktop?

Wuying is Alibaba Cloud's desktop-as-a-service (DaaS) platform that delivers full-featured Windows and Linux desktops from the cloud. Users access their desktop through a lightweight client on any device — laptop, tablet, thin client, or even a web browser. All processing happens on Alibaba Cloud's infrastructure, with only screen updates and input signals transmitted over the network.

2. Wuying Product Lineup 2026

Edition	vCPU / RAM	GPU	Monthly Cost	Target Use Case
Basic Office	2 vCPU / 4 GB	—	~$18/mo	Email, docs, light browsing
Standard Office	4 vCPU / 8 GB	—	~$32/mo	General office, productivity apps
Performance	8 vCPU / 16 GB	—	~$58/mo	Developers, finance analysts
GPU Standard	4 vCPU / 16 GB	Grid T4	~$85/mo	Design, video editing
GPU Pro (GWS)	8 vCPU / 32 GB	A10 / L4	~$180/mo	3D rendering, AI workloads
Agentic Computer	8+ vCPU / 32+ GB	GPU-accelerated	Custom	AI agent workflows, automation

Note: Pricing varies by region and configuration. Wuying offers flexible billing: pay-as-you-go, monthly subscription, and annual plans with up to 30% savings. Channel partners provide additional discounts.

3. Hands-On Performance Review

3.1 Office Productivity

The 4 vCPU / 8 GB standard configuration handles Microsoft Office, browser-based work, and collaboration tools (Zoom, Teams, Slack) without breaking a sweat. Boot time is under 30 seconds, and application launches are near-instant thanks to SSD-backed storage.

3.2 Development Workloads

For software developers, the 8 vCPU / 16 GB configuration with VS Code, Docker, and browser-based testing tools performs comparably to a mid-range laptop. The key advantage? Your entire dev environment is always on, always synced, and accessible from anywhere.

3.3 GPU & Design Workloads

The GPU-accelerated editions handle Photoshop, Illustrator, and even basic 3D modeling in Blender smoothly. The GWS GPU Workstation series with professional-grade GPUs is suitable for architectural visualization, game development, and video production.

3.4 AI Workloads

With the new Agentic Computer tier, Wuying becomes more than just a desktop — it's an AI-powered workstation. Run local LLMs, automate workflows with AI agents, and process large datasets without investing in expensive GPU hardware.

4. TCO Comparison: Wuying vs. Traditional PCs

Let's compare the 3-year total cost of ownership for a 100-person organization:

Cost Component	Traditional PCs (100 users)	Wuying Standard (100 users)	Savings
Hardware purchase	$80,000 ($800/PC)	$10,000 (thin clients)	$70,000
Software licenses	$24,000	$24,000	$0
IT administration	$120,000 (2 FTEs)	$60,000 (1 FTE)	$60,000
Hardware refresh (3yr)	$80,000	$0 (cloud-based)	$80,000
Cloud subscription	$0	$115,200 ($96/user/yr)	-$115,200
Security & compliance	$30,000	$6,000 (built-in)	$24,000
3-Year Total	$334,000	$215,200	$118,800 (35.5%)

Note: The TCO calculation shows Wuying delivers approximately 35% savings over 3 years while improving security, manageability, and business continuity. Savings increase with larger deployments and channel partner discounts.

5. Key Advantages of Wuying

Enhanced Security: Data never leaves the cloud. No risk of data theft from lost or stolen devices. Centralized policy enforcement and access controls.
Rapid Scaling: Deploy 10 or 1,000 desktops in minutes. Perfect for seasonal workers, project-based teams, and rapid hiring.
Business Continuity: Desktops are automatically backed up and can be accessed from any device. Office fire, natural disaster, or pandemic — work continues uninterrupted.
Centralized Management: IT teams manage all desktops from a single console. Software updates, patches, and configuration changes happen in one place.
Green IT: Thin clients consume 80%–90% less power than traditional PCs, reducing both electricity costs and carbon footprint.

Calculate Your Wuying Savings

Get a customized TCO analysis for your organization

View Wuying Plans →

GPU Workstation Details →

Who Should (and Shouldn't) Switch to Wuying

Ideal Candidates

Companies with distributed or remote teams
Regulated industries (finance, healthcare, government) needing data security
Organizations with seasonal or fluctuating headcount
Design and creative teams needing GPU-accelerated workstations
Companies pursuing AI agent and automation initiatives

Less Suitable For

Environments with extremely limited or unreliable internet connectivity
Workloads requiring specialized hardware not available in the cloud
Very small teams (1–3 users) where the management overhead is minimal

Frequently Asked Questions

What internet speed do I need for Wuying?

For standard office use, 5–10 Mbps per user is sufficient. For GPU-accelerated desktops with video or 3D, 20–50 Mbps per user is recommended. Wuying's adaptive streaming technology adjusts quality based on available bandwidth.

Can I use my existing software licenses?

Yes. Wuying supports bring-your-own-license (BYOL) for most software. You can also use Alibaba Cloud's marketplace for pre-licensed applications. Consult your channel advisor for licensing strategy.

Is Wuying secure enough for sensitive data?

Wuying is widely used in financial services, healthcare, and government. It supports VPC isolation, data encryption at rest and in transit, multi-factor authentication, and granular access controls. It meets major compliance standards including ISO 27001, SOC 2, and GDPR.

How does Wuying compare to Citrix or VMware Horizon?

Unlike traditional VDI solutions, Wuying is a fully managed service — you don't need to deploy and maintain VDI infrastructure. This eliminates upfront capital expenditure and reduces ongoing management complexity by 60%–70%.

<div

Golden Arowana & FAW Toyota Join Qwen Office on the Same Day: Enterprise AI Enters Deep Water

liekeai — Thu, 24 Sep 2026 04:47:03 +0000

Golden Arowana & FAW Toyota Join Qwen Office on the Same Day; "Enterprise Context" Launched: Enterprise AI Enters Deep Water (Sep 2026) | Lieke Tech

# Golden Arowana & FAW Toyota Join Qwen Office on the Same Day: Enterprise AI Enters Deep Water

At the Yunqi Conference in Hangzhou on Sept 22, major enterprises including Yihai Kerry's Golden Arowana and FAW Toyota announced adoption of Qwen Office; the same day, Qwen Office launched "Enterprise Context", a data management product for enterprise-AI collaboration — a clear step from general AI office tools toward deep enterprise customization.
✅ Golden Arowana / FAW Toyota
✅ "Enterprise Context" product
🚀 After 30M Qwen Office users

Explore Alibaba Cloud ECS →
Claim SMB Coupons

📌 30-Second Summary

What happened: On Sept 22, at the Yunqi Conference in Hangzhou, Yihai Kerry's Golden Arowana (a leading consumer food brand) and FAW Toyota (a leading joint-venture automaker) announced adoption of Qwen Office. The same day, Qwen Office launched "Enterprise Context", a product that manages how enterprise data becomes AI context — the missing middle layer for enterprise-grade AI office.

Why it matters: Enterprise AI office is moving from "general assistants" to "deep organizational systems". Getting company data — documents, systems, knowledge bases — safely and accurately into AI context has been the last barrier to enterprise-scale adoption.
Acceleration in one month: 30M users in early September → CIMC Enric group-wide adoption on Sept 7 → Golden Arowana & FAW Toyota on Sept 22. Consumer and automotive leaders are now publicly standing behind the platform.
What it means for you: Enterprise AI office has entered the selection phase. For both procurement decision-makers and technical teams, now is the window to evaluate enterprise AI office tools and their underlying LLM platform.

🏭 At Yunqi: Why Leading Enterprises Are Standing Up

The enterprise customer lineup around Qwen Office visibly upgraded at the Yunqi Conference: Golden Arowana — a household consumer brand in cooking oil and food, reaching hundreds of millions of families; FAW Toyota — a major joint-venture automaker with a large national dealer and supply-chain network. Combined with CIMC Enric (clean-energy equipment, business in 100+ countries), Qwen Office's enterprise references now span energy, manufacturing, consumer goods and automotive.

Adoption by industry leaders is a stronger signal than any keynote: enterprises of this size must pass four gates before buying an AI office product — data security, compliance, stability and integrability. Winning consumer and automotive giants at the same time means the product has cleared the enterprise bar, not just the "personal assistant" bar.

Enterprise adoption timeline (September 2026)

Early Sept: Qwen Office passes 30M users in 30 days; more than half are enterprise users
Sept 7: CIMC Enric signs a strategic agreement — group-wide adoption (R&D, procurement, production, sales, service + internal management), integrating SAP/Salesforce-class systems
Sept 22: Yunqi Conference — Golden Arowana, FAW Toyota and other large enterprises announce adoption; "Enterprise Context" launches

Earlier, companies including Changan Automobile, Huifu, Transfar Group and Laoxiangji had already onboarded. From manufacturing to consumer brands, from SOEs to private enterprises, adoption is shifting from individual lighthouse accounts to industry-wide rollout.

🧠 What Is "Enterprise Context" and Why It Is the Key Missing Layer

The biggest pain point in enterprise AI office is not model capability — it's that models don't "know" your company. Your policies, product docs, customer data and project conclusions live in different systems. Without that context, a general chat model gives generic answers; enterprises need answers grounded in their own data.

"Enterprise Context" is positioned as a data management product for enterprise-AI collaboration. It addresses three things:

Layer	Problem with traditional approaches	Enterprise Context approach
Data access	Docs, sheets and system data are scattered; humans manually prepare content for AI	Manage enterprise data sources uniformly; AI reads within controlled boundaries
Context management	AI conversations are fragmented; no knowledge is retained across sessions or teams	Persist enterprise knowledge, business rules and past decisions as reusable context
Security boundaries	Employees paste sensitive data to AI with no governance	Enterprise-grade permissions and compliance; who sees what context is managed and auditable

In short, "Enterprise Context" fills the middle layer that turns AI office from a personal productivity tool into an organizational system. Without it, AI writes emails and makes spreadsheets. With it, AI behaves like a colleague who actually knows the company — answering from real company data and executing tasks that match company workflows.

Why it appeals to large enterprises: A consumer company like Golden Arowana has knowledge spanning channels, distributors, marketing and quality control; a carmaker like FAW Toyota spans R&D, supply chain and after-sales. Making AI "know" this is a direct source of cost reduction and a step toward treating data as an asset. Enterprise Context turns "AI using company data" from an ungovernable gray zone into a governable enterprise capability.

📈 Three Signals: Enterprise AI Office Enters the Selection Phase

Put the three September milestones together and the direction of the market is clear:

Signal 1: From personal trials to organizational procurement. 30M users validated consumer and SMB penetration; deals like CIMC Enric, Golden Arowana and FAW Toyota validate organizational buying. As leaders enter, enterprise AI office joins CIO/CTO selection lists.
Signal 2: From general assistants to industry depth. "Enterprise Context" shows the product evolving from general office capability toward industry know-how — first data access, then industry solutions.
Signal 3: From point tools to platform ecosystems. Qwen Office sits on Tongyi models plus the Alibaba Cloud Model Studio platform — model, cloud foundation and enterprise services in one stack. Choosing AI office is increasingly choosing a technology stack you can evolve with.

🏢 How to Evaluate Enterprise AI Office: A Practical Checklist

If your organization is evaluating AI office or enterprise AI, use this order:

Data access first: Can it connect to your documents, spreadsheets and business systems? Breadth and depth of data access set the ceiling for how much AI "knows" your business.
Then context management: Can enterprise knowledge be persisted, reused and shared across teams? Can different roles and permissions see different scopes of context?
Security & compliance: Where is data stored? Are access controls and audit trails enterprise-grade? Is sensitive data governed?
Foundation & ecosystem: Are models continuously updated? Is the platform stable? Is it deeply integrated with a major cloud vendor for smooth upgrades?

For SMBs that want a fast validation, start with the free/team tier: put one real workflow into AI office (customer records, policy Q&A, weekly report summaries), then evaluate the enterprise tier. LLM capabilities are exposed through the Alibaba Cloud Model Studio stack, with up to 1M free tokens per model for new users — enough for a low-cost technical proof of concept.

☁️ For Developers and Technical Teams: The Foundation Is Model Studio

Qwen Office is the product-layer packaging of Alibaba Cloud's LLM capabilities. If your team wants to build "Enterprise Context"-style features into your own systems — knowledge-base Q&A, RAG over enterprise data, internal agents — the foundation is Alibaba Cloud Model Studio:

Broad model choice: One API key for Qwen3.8 series, DeepSeek, MiniMax and more, OpenAI/Anthropic-compatible interfaces;
Controlled cost: Qwen3.8-Flash input from ~$0.11/M tokens; official tests show ~100% speedup and ~75% fewer tokens on typical office tasks;
New-user credits: Up to 1M free tokens per model (90 days), plus 50% off selected models 22:00–08:00;
Hosting: For teams deploying their own agents, Alibaba Cloud overseas nodes (Singapore/Hong Kong) need no ICP filing; new users get free trials plus voucher bundles.

Prices and promotions follow the latest Alibaba Cloud official pages; new-user benefits require real-name verification.

⏰ Alibaba Cloud September Deals · Exclusive Channel

Lightweight cloud servers from ~20% of list price for new customers, from around $14/year
Model Studio: up to 1M free tokens per model (90 days) for new users; 50% off selected models 22:00–08:00
Overseas nodes: free trial + voucher bundles, no ICP filing required
Enterprise procurement: annual ECS and GPU servers at volume discounts via channel pricing

🔥 Claim Coupons Now →
Explore ECS & Model Studio →

Offers subject to the latest Alibaba Cloud official pages; new-user benefits require real-name verification

❓ FAQ

Is "Enterprise Context" just an enterprise knowledge base?

Related direction, but more systematic. Enterprise Context manages the full loop of how enterprise data becomes AI context — data access, knowledge persistence, permission boundaries and cross-session reuse. Think of it as the capability layer that makes AI "understand the company", not just a Q&A box over documents.

Can SMBs use it?

Yes. Qwen Office offers a free personal tier. SMBs can start on the team tier, validate with a real workflow (policy Q&A, customer records, weekly reports), then consider the enterprise tier for security, compliance, permission layering and usage governance at scale.

How is data security handled?

Adoption by large enterprises such as CIMC Enric (manufacturing), Golden Arowana (consumer) and FAW Toyota (automotive) is itself a validation of enterprise-grade security. For sensitive data scenarios, use the enterprise tier and rely on the official security whitepaper and contract terms.

How do we get started with Qwen Office / Model Studio?

Individuals and small teams can register and try it directly. For enterprise procurement, annual resources, GPU servers or Model Studio enterprise onboarding, use our Alibaba Cloud channel for the latest campaigns and channel pricing: SMB coupon campaign ｜ ECS & Model Studio.

What should I do now?

If you're a decision-maker: add enterprise AI office to your shortlist and evaluate on data access, context management, security/compliance and foundation ecosystem. If you're a technical team: use the free token credits to run one knowledge-base Q&A or agent proof of concept — the smallest possible test of whether this path works in your scenario.

Try It Now: Make AI "Know" Your Enterprise

Qwen Office free tier · Up to 1M free tokens per model on Model Studio · Cloud servers from ~20% off for new customers

Explore Alibaba Cloud →
Claim SMB Coupons

Sources: public announcements from the 2026 Yunqi Conference (Hangzhou), official Qwen Office releases, and public media reports (Sept 22, 2026); product features and prices follow the latest official Alibaba Cloud pages.

Website Builder Guide 2026: 4 Paths to Building Your Business Website — Compared

liekeai — Wed, 23 Sep 2026 04:47:02 +0000

Every business needs a website, but the "right" way to build one depends on your goals, budget, technical skills, and growth plans. In 2026, there are four main paths to getting a business website online — each with distinct trade-offs in cost, flexibility, and effort. This guide compares all four approaches and helps you choose the best one for your situation.

Build Your Professional Website Today

Alibaba Cloud Wanwang: all-in-one website building with domain, hosting, and templates

Explore Website Builder →

View All Offers →

The Four Website Building Paths

Path 1: DIY Website Builders (No-Code)

Platforms like Alibaba Cloud Wanwang Website Builder provide drag-and-drop interfaces with pre-designed templates. No coding required — you pick a template, customize text and images, and publish.

Path 2: WordPress / CMS Platforms

WordPress powers over 40% of all websites. It offers a vast ecosystem of themes and plugins, giving you more flexibility than no-code builders but requiring more technical know-how and maintenance.

Path 3: Custom Web Development

Hire developers to build a fully custom website using modern frameworks (React, Vue, Next.js). Maximum flexibility and performance, but highest cost and longest timeline.

Path 4: ECS Self-Hosted + Open-Source

Deploy open-source platforms (WordPress, Ghost, Drupal, etc.) on your own ECS server. Full control, scalable, and cost-effective for those comfortable with server management.

2. Side-by-Side Comparison

Factor	DIY Builder	WordPress (Managed)	Custom Development	ECS Self-Hosted
Initial Cost	$5 – $50/mo	$10 – $100/mo	$5,000 – $50,000+	$10 – $200/mo + setup
Time to Launch	1–7 days	1–4 weeks	2–6 months	1–4 weeks
Technical Skill Required	Very low	Low-Medium	High	Medium-High
Design Flexibility	Limited to templates	High (themes + plugins)	Unlimited	High (depends on platform)
Scalability	Limited	Good (with hosting plan)	Excellent	Excellent
Maintenance Effort	None (managed)	Low-Medium	Ongoing dev team	Medium-High
SEO Capability	Basic-Good	Excellent (Yoast, etc.)	Depends on implementation	Excellent
Best For	Small business, solo	Most businesses	Tech startups, enterprise	Tech-savvy businesses

3. Deep Dive: Each Path Explained

3.1 DIY Website Builders

DIY website builders are the fastest and easiest way to get online. Alibaba Cloud Wanwang offers a comprehensive website builder with hundreds of professional templates, built-in SEO tools, domain registration, and reliable hosting — all in one package.

Pros:

Launch in hours, not weeks
No technical skills needed
All-inclusive (hosting, domain, SSL, email)
Predictable monthly cost
Mobile-responsive templates

Cons:

Limited design customization beyond templates
Can't easily export your site to another platform
Less SEO control than WordPress
E-commerce features may be limited on lower tiers

Best for: Local businesses, freelancers, startups needing a quick professional web presence.

Try Wanwang Website Builder

Professional templates, built-in SEO, all-in-one pricing

Start Building →

Browse All Offers →

3.2 WordPress / Managed CMS

WordPress dominates the CMS market for good reason. Its massive plugin ecosystem means you can add almost any functionality — e-commerce, memberships, booking systems, SEO tools — without coding. Managed WordPress hosting takes care of server maintenance, updates, and security.

Pros:

Massive theme and plugin ecosystem
Excellent SEO capabilities
Large talent pool of WordPress developers
Can grow with your business
E-commerce with WooCommerce

Cons:

Requires ongoing maintenance (updates, backups, security)
Can be slow without optimization
Plugin conflicts and security vulnerabilities are common
Customization beyond theme options needs coding

Best for: Content-heavy sites, blogs, small-to-medium e-commerce, businesses wanting marketing control.

3.3 Custom Web Development

For businesses with unique requirements or ambitious growth plans, custom development using modern frameworks delivers the best performance, security, and user experience. You have complete control over every aspect of the site.

Pros:

100% customization — exactly what you need
Best performance and page speed
Clean codebase without bloat
Scalable architecture from day one
Unique brand identity

Cons:

High upfront cost
Long development timeline
Requires ongoing development resources
Higher risk if requirements aren't clear

Best for: Tech startups, enterprise applications, SaaS products, high-traffic e-commerce.

3.4 ECS Self-Hosted

Deploy any open-source platform on Alibaba Cloud ECS for maximum control and scalability. You get root access to your server, can install any software, and scale resources as needed. This is the sweet spot for businesses that want flexibility without paying premium managed hosting prices.

Pros:

Full server control and root access
Install any software stack
Scalable — upgrade/downgrade anytime
Cost-effective at scale
Integrates with other Alibaba Cloud services

Cons:

Requires server administration skills
Responsible for security, updates, backups
Steeper learning curve
Time-consuming setup and maintenance

Best for: Tech-savvy businesses, developers, agencies managing multiple client sites.

4. Decision Framework: Which Path Is Right for You?

Ask Yourself These Questions:

What's my budget? Under $100/mo → DIY builder or basic WordPress. $100–$500/mo → Managed WordPress or ECS self-hosted. $5K+ upfront → Custom development.
How soon do I need to launch? Under 2 weeks → DIY builder. 1–2 months → WordPress. 3+ months → Custom.
How technically skilled is my team? No technical skills → DIY builder. Some technical comfort → WordPress. Dev team available → Custom or ECS.
What features do I need? Simple brochure site → DIY. Blog + marketing → WordPress. Web app / SaaS → Custom development.
How much traffic am I expecting? Under 10K visits/mo → Any option. 10K–100K → Managed WordPress or ECS. 100K+ → ECS with CDN or custom architecture.

5. Hosting Your Website on Alibaba Cloud

Regardless of which building path you choose, Alibaba Cloud provides reliable, scalable hosting infrastructure. Here are the recommended hosting approaches:

Website Type	Recommended Hosting	Starting Price
Simple brochure / DIY site	Wanwang Website Builder	~$5/mo
WordPress blog / small business	Basic ECS + OSS + CDN	~$10/mo
E-commerce / high-traffic	ECS + RDS + OSS + CDN + WAF	~$50/mo
Enterprise / web app	Multi-AZ ECS + SLB + RDS + Redis	~$200/mo

Note: New to Alibaba Cloud? Start with a {a_tag(0, 'new user special offer')} to get ECS at deeply discounted rates for the first year. Channel partners provide additional discounts on top of promotional pricing.

6. Essential Website Features Every Business Needs

Regardless of which path you choose, make sure your website includes:

SSL Certificate: HTTPS is essential for security and SEO (free Let's Encrypt or Alibaba Cloud SSL)
Mobile Responsiveness: 60%+ of traffic is mobile — test on real devices
SEO Fundamentals: Title tags, meta descriptions, structured data, fast load times
Analytics: Google Analytics or similar to track performance
Contact Form: Make it easy for customers to reach you
Clear CTA: What do you want visitors to do? Make it obvious
Backup System: Automated backups protect against data loss
CDN: Speed up global delivery with Alibaba Cloud CDN

Frequently Asked Questions

Can I start with a DIY builder and upgrade later?

Yes, but with caveats. Most DIY builders are proprietary platforms — you can't easily export the design. You'll need to rebuild on the new platform. However, content (text, images) can be migrated manually. For this reason, many businesses start with WordPress for better long-term flexibility.

Do I need a domain name separately?

Most website builders and hosting providers offer domain registration as an add-on. Wanwang includes domain options as part of their all-in-one packages. You can also register a domain separately and point it to any hosting provider.

How much does it really cost to run a business website?

Basic DIY: $5–$20/mo. WordPress with hosting: $15–$100/mo. ECS self-hosted: $10–$200+/mo depending on resources. Custom development: $5K–$50K+ upfront plus $500–$5K/mo maintenance. Don't forget marketing budget, which usually exceeds hosting costs!

Is WordPress still relevant in 2026?

Absolutely. WordPress continues to evolve with the block editor (Gutenberg), full-site editing, and headless CMS options. It powers 40%+ of the web for good reason — it's flexible, well-supported, and has an unmatched ecosystem. Modern alternatives exist, but none match WordPress's combination of maturity, flexibility, and community.

<div

Tencent WorkBuddy Review 2026

liekeai — Tue, 22 Sep 2026 04:47:03 +0000

1. What Is WorkBuddy?

WorkBuddy (Chinese name: "沃克巴迪") is an all-in-one AI office agent launched by Tencent Cloud's CodeBuddy team in March 2026. It's not a chatbot — it's a desktop application that can directly operate your computer and complete multi-step tasks autonomously.

In plain terms: you say "turn this Excel data into a PowerPoint," and WorkBuddy opens Excel, reads the data, analyzes it, generates the slides, and saves the file — all without manual intervention.

Key stats: 8M+ MAU within 2 months of launch; 16.5% adoption among non-technical staff; 100+ domain experts built in; 70,000+ Skills available; supports Windows, Mac, iOS, Android, and HarmonyOS.

2. Core Features

Capability	Description
Natural Language PC Control	Issue commands in plain language; WorkBuddy breaks them down and executes multi-step operations locally
Cross-App Automation	Simulates human switching between Word, Excel, PowerPoint, and browsers to complete complex workflows
Multi-Format File Processing	Read, write, merge, and convert Word, Excel, PPT, PDF, images, and more
Multi-Model Switching	Built-in Hunyuan, DeepSeek, GLM, Kimi, MiniMax; supports private model integration
SkillHub Marketplace	200+ pre-built skill packs, 70,000+ Skills, compatible with OpenClaw skills
Remote Control	Send commands via WeCom/QQ/Lark/DingTalk; let your PC work while you're away
100+ Domain Experts	Pre-configured specialists for recruiting, research, frontend dev, legal, marketing, and more
Cloud Persistence	Tasks continue running in the cloud even after the client is closed

3. Pros and Cons

✅ Advantages

Zero deployment: Download, install, and use — no coding required
Gets real work done: Actually manipulates files and software, not just chat
Tencent ecosystem: Deep WeCom/QQ integration, easy remote control
Multi-model: Switch between Hunyuan, DeepSeek, GLM, and others
Massive skill library: 70,000+ Skills covering most office scenarios
Cross-platform: Windows, Mac, iOS, Android, HarmonyOS

❌ Drawbacks

Opaque Credits pricing: Charged by task complexity; hard to predict costs
Requires PC on: Desktop app depends on local machine running
Privacy concerns: Requires folder access permissions
Complex tasks may fail: Multi-step workflows occasionally break or produce errors
Enterprise pricing adds up: Per-seat subscription; long-term costs need calculation
Vendor lock-in: Deep Tencent ecosystem integration makes migration costly

4. Pricing: How Credits Work

WorkBuddy uses a Credits system. New users receive 5,000 free Credits. Enterprise plans include 2,000 base Credits + 2,000 bonus Credits per seat per month.

Plan	Price	Credits	Best For
Personal	Free + in-app purchases	5,000 on signup	Individual users trying it out
Enterprise SaaS	Per seat / month	2,000 base + 2,000 bonus	Small-to-medium teams
Enterprise Private	Contact sales	Custom	Finance, government, regulated industries

⚠️ Important: Credits are consumed based on task complexity. Generating a PowerPoint can cost hundreds of Credits; a deep research report may cost over a thousand. The 5,000 free Credits may last only 1–2 weeks for heavy users. Actual monthly spend depends on usage intensity — power users could spend hundreds of yuan per month.

5. WorkBuddy vs. DIY on Alibaba Cloud

If you're a developer, SMB, or technical team, instead of paying per Credit, consider building your own AI workflow on Alibaba Cloud. Set it up once and run it long-term at a fraction of the cost.

Dimension	Tencent WorkBuddy	Alibaba Cloud DIY
Startup Cost	5,000 free Credits	ECS from $4.50/mo + 70M free Tokens
AI Model	Switch between Hunyuan/DeepSeek/GLM	Qwen3.5-Flash from $0.028/M tokens — among the cheapest globally
Runtime	Requires local PC powered on	Cloud ECS running 24/7
Customization	Limited to Skill ecosystem	Fully autonomous, any workflow you can code
File Processing	Native desktop ops, out of the box	Python scripts + APIs; requires some dev skill
Monthly Cost (Heavy)	Hundreds in Credits	ECS $4.50/mo + API pay-as-you-go, typically

💡 Cost calculation: Processing 100 documents/reports per month — WorkBuddy Credits might consume 5,000–10,000 Credits (roughly $14–42/mo). On Alibaba Cloud: ECS at ~$4.50/mo + Qwen3.5-Flash API at ~$1.50–4/mo. Total under $7/mo, running 24/7 without depending on a local PC.

6. Quick Start: DIY AI Workflow on Alibaba Cloud

Step 1: Claim Free Resources

ECS Cloud Server: 2 vCPU / 2GB RAM from ~$4.50/mo (renewal at same price), enough for automation scripts and lightweight apps
ModelStudio (Bailian): New users get 70 million free Tokens + 100 AI images + 50 seconds of video, valid for 180 days
International: $200 cloud credit + 70M AI Tokens, 30-day refund guarantee

Step 2: Choose Your AI Model

Alibaba Cloud's ModelStudio offers the full Qwen model family:

Qwen3.5-Flash: From $0.028/M input, $0.28/M output tokens, 1M context window, multimodal — top choice for daily tasks
qwen-turbo: $0.042/M input, $0.084/M output, 130K context, excellent cost-performance
Qwen3.5-Plus: $0.11/M input, $0.67/M output, enhanced reasoning

Step 3: Build Your Automation

Deploy Python scripts on ECS that call ModelStudio APIs to achieve: automated document generation, data analysis, batch file processing, scheduled report delivery, web monitoring — everything WorkBuddy does, but fully under your control at lower cost.

7. Who Should Choose What?

Choose WorkBuddy if: You're in a non-technical role (HR, finance, operations, marketing), don't want to code, need quick onboarding, primarily do document processing and PPT generation, and your budget allows per-Credit payments.

Choose Alibaba Cloud DIY if: You're a developer or technical team, need 24/7 automation, have custom workflow requirements, care about long-term costs, or want full data control. After a one-time setup, monthly costs can be as low as $4.50 (ECS) + pay-as-you-go API.

Best of both worlds: Non-technical teams can use WorkBuddy for daily office tasks while technical teams build backend automation on Alibaba Cloud ECS + ModelStudio — the two complement each other. But if you must choose one, the DIY approach wins on long-term ROI.

8. FAQ

Is WorkBuddy free?

The personal edition is free to download with 5,000 signup Credits. After that, you'll need to purchase Credits or subscribe to the enterprise plan. 5,000 Credits last light users about 1–2 weeks; heavy users may exhaust them in days.

Is WorkBuddy secure?

WorkBuddy uses sandbox isolation, skill standardization, and dangerous operation interception as multi-layer defenses. Enterprise edition supports SSO, IP whitelisting, and audit trails. However, as a desktop agent it does require folder access — we recommend authorizing only necessary directories.

Can I build a DIY solution without coding?

ModelStudio provides visual workflows and extensive API documentation. Combined with community tutorials, even beginners can get started quickly. ECS offers pre-configured images for one-click deployment. You can also start with ModelStudio's online experience for the simplest approach.

Qwen3.5-Flash vs DeepSeek — which is better?

After DeepSeek's August 2026 price increase, Qwen3.5-Flash has a clear value advantage: input from $0.028/M (vs. DeepSeek V4 Pro at ~$1.25/M off-peak), output from $0.28/M (vs. DeepSeek off-peak ~$3.75/M), with 1M context and multimodal support. For most daily scenarios, Qwen3.5-Flash is more than sufficient.

Start Building Your AI Workflow Today

ECS from $4.50/mo + 70 million free Tokens. Total cost under $7/mo for heavy usage.

Claim $200 Free Credit →
Browse ECS Instances

Qwen3.8-Flash: Full Review

liekeai — Mon, 21 Sep 2026 04:47:02 +0000

Qwen3.8-Flash Review: $0.15/M Input, 67% Cheaper Than DeepSeek | Lieke Tech

# Qwen3.8-Flash: Full Review

Price cut just one day after launch — $0.15/M input tokens, 89% lower training cost, coding benchmarks match DeepSeek V4 Pro
✅ Input $0.15/M tokens
✅ Output $0.47/M tokens
✅ Cache hit $0.016/M

Claim $200 Free Credit →
ECS from $4.50/mo

📌 Key Takeaways

Launch then price cut: Qwen3.8-Flash launched on August 26 at $0.16/$0.47 per million tokens; Alibaba cut the China-side price on August 27 to ¥0.8 (input) and ¥2.7 (output), with cache hits at ¥0.1/M

Architecture breakthrough: 125B-parameter MoE model activating only 6B per token, plus a 51B N-gram embedding layer — training cost is ~1/9 of Qwen3.7-Plus (89% reduction)
Competitive performance: Scores 62.5 on SWE-bench Pro (agentic coding), leads in 8 of 14 benchmarks, matching DeepSeek-V4-Flash and Claude Opus 4.6 on real-world tasks
Mega context: Natively supports 262K tokens, extends to 1M with YaRN. Multimodal: text, image, and video input
Cache economics: Cache hit at $0.016/M tokens — ideal for agents, RAG, and long conversations

💰 Latest Pricing (Effective August 27, 2026)

Pricing via Alibaba Cloud Model Studio (international) and Bailian platform (China), per million tokens.

International Pricing (USD)

Billing Item	Price (per 1M tokens)
Input	$0.15
Output	$0.47
Cache hit	$0.016

China Pricing (RMB, after Aug 27 cut)

Billing Item	Launch Price (Aug 26)	Current Price (Aug 27+)	Change
Input	¥1.0	¥0.8	↓20%
Output	¥3.0	¥2.7	↓10%
Cache hit	—	¥0.1	—

💡 Real-world cost example: An AI customer service app handling 1M requests/day (avg 2K input + 500 output tokens):

Qwen3.8-Flash: ~$415/day total

DeepSeek V4 Flash (peak): ~$1,260/day

Savings: ~$300,000+ per year

⚔️ Competitive Pricing Comparison

August 2026 pricing for major LLM APIs, USD per million tokens.

Model	Input	Output	Active Params	Context
Qwen3.8-Flash NEW	$0.15	$0.47	6B (MoE)	1M
Qwen3.5-Flash (≤128K)	$0.11	$0.67	—	1M
Qwen3.7-Flash	$0.034	$0.132	—	256K
Qwen3.8-Max	$2.00	$6.00	—	1M
DeepSeek V4 Flash (peak)	$0.42	$1.26	13B	128K
DeepSeek V4 Flash (off-peak 50%)	$0.21	$0.63	13B	128K
DeepSeek V4 Pro (peak)	$1.26	$3.78	—	128K
GLM-5.3-Flash	~$0.15+	~$0.47+	18B (MoE)	—

Key findings:

Qwen3.8-Flash output is 63% cheaper than DeepSeek V4 Flash peak ($0.47 vs $1.26)
Even against DeepSeek's off-peak pricing, Qwen3.8-Flash remains 25% cheaper on output
API price is 1/10 of GLM-5.3, and up to 1/20 during promotional periods
Qwen3.7-Flash is cheaper but lacks the 1M context and multimodal capabilities of 3.8

🏗️ Architecture Deep Dive

Qwen3.8-Flash (codenamed Qwen3.8-Flash-Next during development) is an early preview of the Qwen4 architecture, with four core innovations:

1. Mixture-of-Experts (MoE)

125B main parameters with only 6B activated per token (10 routed experts + 1 shared expert out of 512)
Additional 51B N-gram embedding layer stored in system RAM (not GPU), acting as a massive "phrase dictionary"
4B multi-token prediction parameters bring total footprint to ~180B

2. QSA Sparse Attention + GDN Hybrid

Qwen Sparse Attention (QSA) works alongside GDN: GDN compresses historical context while QSA selects key information. In high-cache-hit 1M-token scenarios, this delivers 7.6x faster prefill and 4.9x faster decode.

3. Gated Residual Mechanism

Splits the traditional single residual pathway into 4 parallel branches, dynamically gating information flow for better cross-layer communication and training stability.

4. Muon Optimizer

Uses a refined Muon + AdamW hybrid strategy with refitted scaling laws, eliminating batch warm-up and significantly improving convergence efficiency and training throughput.

Result: Training costs are approximately 1/9 of Qwen3.7-Plus (397B params, 17B activated) — an 89% reduction — while delivering superior coding and office task performance.

📊 Benchmark Performance

Per Qwen's official technical report, Qwen3.8-Flash excels across multiple evaluation suites:

Benchmark	Domain	Performance
SWE-bench Pro	Agentic coding	62.5, leading peers
CoWorkBench	Long-horizon office tasks	Beats DeepSeek V4 Flash
Toolathlon Verified	Real-world tool use	Matches Claude Opus 4.6
MathVision	Visual math reasoning	Strong multimodal
AndroidWorld	Mobile agent tasks	Embodied intelligence
ERQA	Embodied reasoning	Multimodal understanding

Across 14 evaluations, the base model (6B activated) achieved the best results in 8. The fine-tuned version shows even stronger performance in coding, agents, and multimodal tasks.

🎯 Best Use Cases

Qwen3.8-Flash is ideal for:

AI coding assistants: 62.5 on SWE-bench Pro, repository-level understanding with 1M context
AI agents / tool use: Native parallel tool calls, extremely low cost for long-horizon tasks
Customer service / chatbots: $0.016/M cache hit makes multi-turn conversations nearly free
Document analysis / RAG: 1M tokens processes hundreds of pages, cache speeds up 8x
Multimodal apps: Text + image + video input, visual math and chart analysis
High-volume batch processing: 5,000 RPM / 5M TPM rate limits for enterprise scale

🚀 How to Get Started

Option 1: Alibaba Cloud Model Studio (International)

Access Qwen3.8-Flash via Model Studio with OpenAI-compatible API. New users get free credit. The model serves on QwenCloud with 1M context by default.

Claim $200 Free Credit →

Option 2: Bailian Platform (China)

Available on Alibaba Cloud's Bailian platform with new user free token quota. First to receive the latest Qwen releases.

Bailian Console →

Option 3: Open Weights (Self-hosted)

The open-weight version, Qwen3.8-Flash-Next, is available on Hugging Face and ModelScope for download, fine-tuning, and local deployment. The production version with built-in tools and 1M default context is served via QwenCloud API.

New user offers:

Alibaba Cloud international: up to $200 free trial credit
Bailian platform: free token quota for new users (90 days)
ECS cloud servers: from ~$4.50/month (2 vCPU, 2GB RAM)

📰 September 19, 2026 update: Qwen3.8-27B generates working web apps from a single prompt

Qwen3.8-27B keeps trending in developer circles this week. In a hands-on test published by QbitAI on September 19, a single prompt led Qwen3.8-27B to produce a 52KB offline data-analysis tool in about 5 minutes — automatically cleaning tables, aggregating revenue stats and rendering bar charts — and even a realistic 12306 ticket-booking page (a convincing front-end shell that, by default, does not connect to live data or payment backends). Engineer Alok paired Qwen3.8-27B with Cerebras to build an offline "AI computer desktop" that regenerates historical versions of Google, YouTube and other sites on the fly at roughly 1,950 tokens/s (individual benchmark). On September 18, the Qwen team also launched Qwen3.8-Omni-Flash, a next-generation omnimodal model with text/image/audio/video inputs and a 1M-token context, cutting per-hour audio-input API pricing by more than 98% versus the previous generation. From local 27B deployment to cloud APIs, the Qwen3.8 family is making "one-prompt tool delivery" real — for enterprises, calling Qwen3.8-Flash, Qwen3.8-27B and Omni-Flash via Alibaba Cloud Model Studio / Bailian is the most cost-effective way to try this generation of models.

📰 September 8, 2026 update: Qwen3.8-27B quantization benchmark

This week a Hacker News benchmark post by Quesma CTO Piotr Migdał gave a definitive answer for running Qwen3.8-27B locally: the 17GB Q4_K_M 4-bit quant matches the 55GB BF16 full model on Terminal-Bench 2.1 and fits on a 24GB consumer card (e.g. RTX 4090) with room for ~64K tokens of context. At 4-bit, GPQA Diamond, IFBench and Terminal-Bench 2.1 scores are statistically indistinguishable from the reference, at one-third the on-disk size. Compression hits a clear cliff at 1-bit: the 6.2GB UD-IQ1_S falls to near-random on GPQA Diamond, and the xhigh reasoning effort actually makes it worse (the model often runs out of token budget without committing to an answer). Unsloth's stated minimum for agentic / tool-calling work remains the 9.8GB Q2_K_XL. For teams weighing a self-hosted 27B deployment, this is the most current, third-party-validated hardware and quantization guidance available.

Source: Quesma: Benchmarking Qwen3.8 27B quantizations (202 HN points, 99 comments on Sept 8, 2026). The original pricing and architecture analysis above remain valid; this card only adds the local-deployment sizing reference.

❓ FAQ

Qwen3.8-Flash vs Qwen3.8-27B — what's the difference?

Qwen3.8-Flash is a MoE model (125B total / 6B active) optimized for cost efficiency and high throughput. Qwen3.8-27B is a dense vision-language model (all 27B active) for deep reasoning at $0.42/$3.08 per million tokens. Choose Flash for scale, 27B for depth.

How does the cache hit pricing work?

When a request hits the context cache (e.g., repeated system prompts, long document prefixes), cached input tokens are billed at $0.016/M instead of $0.15/M — a 90% discount. This dramatically reduces costs for agents, RAG, and multi-turn conversations.

What input modalities are supported?

Qwen3.8-Flash supports text and image input natively (input_modalities: ["text", "image"]), with video analysis in the production version. It handles chart analysis, document OCR, and visual math.

What are the rate limits?

5,000 RPM (requests per minute) and 5,000,000 TPM (tokens per minute), suitable for enterprise-grade high-concurrency deployments.

Start Building with Qwen3.8-Flash Today

$200 free credit + ECS from $4.50/month + open-weight models available

Claim $200 Free Credit →
View ECS Plans

⏰ Alibaba Cloud September Deals · Exclusive Channel

New users: free trial credits + coupon bundle across ECS, databases and more
Model Studio (Bailian) LLM platform: up to 1M free tokens per model for 90 days; up to 50% off for selected models during 22:00-08:00 (UTC+8) off-peak hours
Overseas regions: Singapore / Hong Kong / US ECS with no ICP filing required
Enterprise: annual ECS and GPU instances with up to 40% off, plus channel-only pricing on request

🔥 Claim Free Credits →
View ECS Plans →

Offers subject to official campaign pages; new-user deals require identity verification.

Pricing sources: Alibaba Cloud Bailian announcement (Aug 27, 2026), Qwen official blog, Alibaba Cloud Model Studio pricing page

Qwen3 API Pricing: Complete Guide

liekeai — Sun, 20 Sep 2026 04:47:02 +0000

Latest Update (September 2026)

Beyond text models, the Qwen3 voice ecosystem is heating up: Alibaba Cloud Model Studio (Bailian) now serves official Qwen3-TTS-Flash and Qwen3-ASR-Flash-Realtime endpoints (17 expressive voices, multi-language, emotion recognition, dialect support). Independent inference providers are also pushing the open-weight Qwen3-TTS 1.7B model to the top of Coval's voice AI benchmarks - #1 latency for speech-to-text and #1 word-error-rate for text-to-speech as of mid-September 2026. For developers building voice agents, Qwen voice APIs combine leading open-source quality with low per-character cost, and you can get started with the same free token allowance.

Qwen3 Model Pricing Overview

Alibaba Cloud offers a range of Qwen3 models through DashScope/Model Studio. Prices below are per 1 million tokens, in USD for international access and CNY for China mainland.

Model	Input (per M tokens)	Output (per M tokens)	Context	Free Tier
Qwen Code CLI	FREE	FREE	262K	1,000-2,000 calls/day
Qwen3.8-Max (flagship)	~$1.65 (¥12)	~$4.95 (¥36)	128K	70M tokens (new users)
Qwen3.7-Max	~$0.83 (¥6)*	~$2.48 (¥18)*	128K	70M tokens (new users)
Qwen3-Plus	~$0.55 (¥4)	~$1.65 (¥12)	128K	Free quota available
Qwen3.7-Flash (0-32K ctx)	$0.03 (¥0.225)	$0.13 (¥0.974)	256K	Free quota available
Qwen3.7-Flash (32-256K ctx)	$0.10-$0.21	$0.41-$0.83	256K	Free quota available
Qwen3-Coder-Plus	~$0.83 (¥6)	~$2.48 (¥18)	262K	Via Qwen Code free
Qwen3-Coder 480B (Vertex AI)	$0.22	$1.80	262K	Via Qwen Code free
Qwen3.5-Omni (multimodal)	~$0.25 (¥1.8)	~$2.17 (¥15.8)	128K	1M tokens free

** Qwen3.7-Max prices shown are after the 50% limited-time discount (standard: ¥12/¥36). USD conversions approximate at 7.25 CNY/USD.*

Best Value: Qwen3.7-Flash at $0.03/M input tokens is one of the cheapest production-grade LLMs available globally. For coding, Qwen Code CLI is completely free for up to 2,000 calls per day.

Free Tier Breakdown

Option 1: Qwen Code CLI (Always Free)

Mainland China: 2,000 API calls/day via Qwen OAuth or ModelScope
International: 1,000 API calls/day via OpenRouter
No token limits within each call
60 requests/minute rate limit
Uses Qwen3-Coder 480B model

npx @qwen-code/qwen-code@latest

Option 2: New User Tokens

70 million+ free tokens across Qwen models
Valid for 180 days (extended from 90 days)
100 AI image generation credits
50 seconds of video generation
200 CNY no-threshold coupon
No credit card required (international) / real-name verification (China)

Option 3: OpenRouter Free Models

qwen/qwen3-coder:free — 20 RPM, 200 requests/day
qwen3-32b — 30 RPM, 1,000 requests/day on GitHub Models

API Quick Start

from openai import OpenAI

International endpoint

client = OpenAI(
api_key="your-dashscope-key",
base_url="https://dashscope-intl.aliyuncs.com/compatible-mode/v1"
)

response = client.chat.completions.create(
model="qwen3.7-max",
messages=[{"role": "user", "content": "Explain async/await in Python"}]
)
print(response.choices[0].message.content)

Cost Comparison with Competitors

Model	Input/M tokens	Output/M tokens	Free Tier
Qwen3.7-Flash	$0.03	$0.13	Yes (70M tokens)
DeepSeek V3	$0.14	$0.56	No
GPT-4o Mini	$0.15	$0.60	Limited
Claude Haiku	$0.80	$4.00	No
Gemini Flash	$0.075	$0.30	Limited
Qwen Code (CLI)	FREE	FREE	1,000-2,000 calls/day

Money-Saving Tips

Use Flash models for simple tasks: Qwen3.7-Flash is 90% cheaper than Max and sufficient for classification, summarization, and simple chatbots.
Maximize free tier first: 70M tokens can handle approximately 70,000-100,000 typical API requests before needing to pay.
Use Qwen Code for development: 2,000 free coding calls per day replaces $10-20/month Copilot subscriptions.
Context caching discount: Qwen3.7-Flash offers reduced pricing for cached tokens (¥0.225/M vs ¥0.749/M for 32K context).
Coding Plan Lite: Heavy coders can get unlimited Qwen3-Coder for 7.9 CNY (~$1.10) first month.
Qoder paid plans from $20/month: for AI agent deployments, Qoder Pro bundles 2,000 Credits with concurrent agents.

Start Building for Free

Claim your 70M+ free tokens and $200 cloud credits today.

Get $200 Free Credits →

No credit card required · 180-day AI token validity

→ Full Qwen Code setup guide · Free cloud hosting comparison

RAG in Production: Build a Document Q&A Bot with Qwen + Bailian

liekeai — Sat, 19 Sep 2026 04:47:02 +0000

RAG in Production: Build a Document Q&A Bot with Qwen + Bailian

Everyone talks about RAG (retrieval-augmented generation). Fewer people show you the whole pipeline: how documents become chunks, how chunks become vectors, and how those vectors answer your users' questions. This post builds a production document Q&A bot with Qwen on Alibaba Cloud Bailian — the managed path that avoids running your own vector database.

Why managed RAG for most teams

No vector database to operate (no pgvector tuning, no index rebuilds)
Built-in document parsing: PDF, Word, Markdown handled out of the box
Managed embeddings + retrieval + LLM in one API surface
Scales without you thinking about it

The pipeline at a glance

Documents → Parse → Chunk → Embed → Vector Index
↓
User question → Embed → Retrieve top-k → LLM (Qwen) → Answer + citations

Step 1: Create the knowledge base

In Bailian, create a knowledge base and upload your documents. The console handles parsing and chunking. Practical tips:

Chunk size around 500 tokens works well for most business documents
Keep a small overlap between chunks so sentence context survives the split
Name chunks with source metadata — you will need it for citations

Step 2: Wire retrieval + generation

With the knowledge base ID, the application code is short:

from dashscope import Application

app = Application(
app_id=your_application_id,
api_key=api_key
)

response = app.call(
prompt="Summarize our refund policy in two sentences",
rag_options={"pipeline_ids": [knowledge_base_id]}
)
print(response.output.text)

The response can include retrieved sources, which lets you render citations instead of hallucinating confidently.

Step 3: Production concerns

Permissions: gate the bot behind your own auth layer; the knowledge base may contain sensitive docs
Observability: log every question and its retrieved chunks — debugging a RAG app without this is painful
Evaluation: build a small golden set of question/answer pairs and run it after every knowledge base update
Cost control: cache common questions, use a smaller model for easy intents, and monitor token spend per session

Step 4: When to move off the managed path

Managed RAG is the right default. You only need to roll your own when you have very custom chunking, multi-tenant isolation at scale, or regulatory requirements around data residency that the managed service cannot meet. For those cases, the architecture is the same — you just own more of the pieces.

I keep independent notes on building AI applications with Qwen and current cloud pricing at lieke-ai.com. The official Bailian documentation and free-token campaign live here: Alibaba Cloud coupons.

Qwen3-Coder vs GitHub Copilot vs Cursor

liekeai — Fri, 18 Sep 2026 04:47:02 +0000

Qwen3-Coder vs GitHub Copilot vs Cursor: 2026 AI Coding Showdown | Lieke Tech

# Qwen3-Coder vs GitHub Copilot vs Cursor

🔥 New: Qoder Review — Alibaba's 6M-user agentic workbench with Claude/GPT/Gemini/Qwen. Free + BYOK, paid plans from $20/mo.

The complete 2026 AI coding assistant comparison. Pricing, features, context window, and real-world performance — updated August 2026.

Quick Verdict

Your Situation	Best Choice	Why
Budget-conscious developer	Qwen Code CLI FREE	1,000 requests/day, zero cost, 60/min rate limit
Power user / API-based	Qwen3-Coder BEST VALUE	$0.20/$1.00 per M tokens, 1M context, 480B model
GitHub-centric team	GitHub Copilot Pro	Deep GitHub integration, $10/month
AI-native IDE preference	Cursor Pro	Multi-file editing, $20/month

Pricing Comparison

Tool	Free Tier	Paid Plan	API Price (per M tokens)
Qwen Code CLI	1,000 req/day	$0.99/mo (Coding Plan Lite)	N/A (CLI tool)
Qwen3-Coder API	70M tokens free	Pay-as-you-go (from $0.20/M)	$0.20 / $1.00
Qwen3-Coder Flash	70M tokens free	Pay-as-you-go	$0.20 / $0.97
GitHub Copilot	2,000 completions/mo	$10-$100/mo	Credits-based
Cursor	Limited (2-week trial)	$20/mo (Pro)	$20/mo includes usage
Claude Code	Trial only	$20/mo (Pro)	$3/$15 (Sonnet)

Key insight: Qwen3-Coder API costs 25x less than running GPT-4o through Copilot for equivalent token usage. For a developer processing 10M tokens/month, Qwen3-Coder costs ~$12 vs. $125+ for GPT-4o.

🎁 Get 70 Million Free AI Tokens + $200 Cloud Credit — No Credit Card Needed →

Feature Deep Dive

Qwen3-Coder: The Open-Source Powerhouse

Qwen3-Coder is a dedicated 480-billion-parameter coding model from Alibaba's Qwen team, now available through Alibaba Cloud's international API. It supports a 1-million-token context window — enough to process entire codebases in a single request.

Feature	Details
Model Size	480B parameters (MoE architecture)
Context Window	262K standard, up to 1M via API
Languages	90+ programming languages
Free Tier	70M tokens (180 days) + Qwen Code CLI free forever
API Endpoint	https://dashscope-intl.aliyuncs.com/compatible-mode/v1

GitHub Copilot: The Established Standard

Copilot remains the most widely adopted tool with deep IDE integration (VS Code, JetBrains, Neovim, Xcode). In 2026, it shifted to credit-based billing with AI credits for chat/agent features. Code completions remain unlimited on paid plans.
✓ Best IDE integration ✓ Mature and stable ✗ Credit system can get expensive

Cursor: The AI-Native IDE

Cursor is a full VS Code fork with AI built in. It excels at multi-file refactoring and understands entire project structures. The dual-pool pricing (Auto + API) can surprise heavy users.
✓ Best multi-file editing ✓ Strong agent mode ✗ Requires IDE migration ✗ Premium models cost extra

Head-to-Head: Capabilities

Capability	Qwen3-Coder	Copilot	Cursor
Code Completion	Excellent	Excellent	Good
Multi-File Editing	Good (API)	Limited	Excellent
Context Window	1M tokens	Model-dependent	Project-level
Agent/Autonomous Mode	Via Qwen Code	Yes (Pro+)	Yes
Open Source	Yes (weights available)	No	No
Self-Host Option	Yes (Qwen3-Coder 30B)	No	No
Offline Capability	Yes (via Ollama)	No	No
Monthly Cost (avg user)	$0-$20	$10-$39	$20

Getting Started with Qwen Code (Free Forever)

Qwen Code is Alibaba's free AI coding CLI. It works alongside VS Code, JetBrains, or any editor — no migration required.

Installation

Requires Node.js 20+:
npx @qwen-code/qwen-code@latest

What You Get Free

1,000 requests per day (international) / 2,000 (China)

No token consumption — request-based, not token-based
60 requests per minute rate limit
Powered by Qwen3-Coder model
Terminal-based agent that can read files, run commands, and edit code

Upgrade to Coding Plan Lite

For power users: $0.99 first month for unlimited Qwen3-Coder access. After trial, still significantly cheaper than alternatives.

🚀 Start Building with Qwen3 — 70M Free Tokens (180 Days) →

When Each Tool Wins

🏆 Choose Qwen3-Coder if...

You want the lowest cost without sacrificing capability
You need a massive context window for large codebases
You prefer open-source models or self-hosting
You build AI-powered dev tools and need API access
You want a free CLI coding assistant forever

Choose GitHub Copilot if...

Your team lives in GitHub (PRs, Issues, Actions)
You want zero-setup IDE integration
Enterprise compliance and admin controls matter
You need broad model choice (GPT-5, Claude, Gemini)

Choose Cursor if...

Multi-file refactoring is your daily work
You're comfortable switching to an AI-first IDE
Agentic coding (autonomous multi-step tasks) is priority

Cost Calculator: 10M Tokens/Month

Tool/Model	Input Cost	Output Cost	Est. Monthly
Qwen3.7-Flash	$0.03/M × 8M	$0.13/M × 2M	$0.50
Qwen3-Coder Flash	$0.20/M × 8M	$0.97/M × 2M	$3.54
Qwen3-Coder 480B	$0.30/M × 8M	$1.00/M × 2M	$4.40
DeepSeek V3	$0.14/M × 8M	$0.56/M × 2M	$2.24
GPT-4o (via Copilot)	$2.50/M × 8M	$10/M × 2M	$40.00
Claude Sonnet	$3/M × 8M	$15/M × 2M	$54.00

Assumes 80% input / 20% output token split, typical for coding tasks. Qwen3-Coder 480B is 9x cheaper than GPT-4o and 12x cheaper than Claude Sonnet.

Frequently Asked Questions

Is Qwen Code really free?Yes. Qwen Code CLI offers 1,000 free requests per day internationally (2,000 in China) with no token limits. The free tier doesn't expire and doesn't require a credit card.
Can Qwen3-Coder replace Copilot?For most developers, yes. Qwen3-Coder handles code completion, debugging, refactoring, and code review at a fraction of the cost. The main trade-off is less polished IDE integration compared to Copilot's native plugins.
What is the Qwen3-Coder context window?Qwen3-Coder supports 262K tokens standard and up to 1 million tokens through the Alibaba Cloud API. This means it can process entire large codebases in a single conversation.
How does the 70M free token trial work?New Alibaba Cloud international users receive 70 million AI tokens valid for 180 days, plus $200 in cloud credits, 2,000 image generations, and 1,650 seconds of video generation. No credit card required.
Can I self-host Qwen3-Coder?Yes. The Qwen3-Coder 30B model is open-source and can run locally via Ollama (ollama run qwen3:7b for the smaller variant) or deployed on your own GPU infrastructure.

🎁 Claim Your Free 70M Tokens + $200 Credit →

⏰ Alibaba Cloud September Deals · Exclusive Channel

New users: free trial credits + coupon bundle across ECS, databases and more
Model Studio (Bailian) LLM platform: up to 1M free tokens per model for 90 days; up to 50% off for selected models during 22:00-08:00 (UTC+8) off-peak hours
Overseas regions: Singapore / Hong Kong / US ECS with no ICP filing required
Enterprise: annual ECS and GPU instances with up to 40% off, plus channel-only pricing on request

🔥 Claim Free Credits →
View ECS Plans →

Offers subject to official campaign pages; new-user deals require identity verification.