Model Studio Token Plan for Individual: One Subscription for Every AI Model, Up to 3x More Value

Build more, spend less — one plan, every modality. The Model Studio Token Plan for Individual is now live — featuring the exclusive debut of Qwen3.8-Max-Preview, plus access to every text, video, image, and audio model on the platform at up to 3x the usage compared to pay-as-you-go pricing. Early-bird launch pricing starts at just $6/month — and if you're already using off-peak rates, those discounts stack with your Token Plan credits.

Whether you're using Qwen3.8-Max-Preview for full-stack development and data analysis, generating video content with HappyHorse 1.1, or prototyping multimodal applications, the Token Plan eliminates per-call cost anxiety and lets you focus on shipping.

Qwen3.8-Max-Preview — First Available on Token Plan

Qwen3.8-Max-Preview debuts exclusively on the Model Studio Token Plan for Individual — the latest and most capable model in the Qwen family, available here before anywhere else.

Whether you're doing full-stack development, data analysis, or automating Office workflows, Qwen3.8-Max-Preview handles it with ease. It combines frontier-level text reasoning with vision understanding, making it equally capable of reading a complex codebase, interpreting a spreadsheet screenshot, or drafting a slide deck outline from a rough sketch.

What Qwen3.8-Max-Preview Excels At

• Full-stack development — multi-file code generation, refactoring, debugging, and architecture reasoning across frontend, backend, and infrastructure
• Data analysis — interpreting charts, processing CSVs, writing SQL, and building dashboards from natural language instructions
• Office productivity workflows — document drafting, email automation, presentation structuring, and meeting summarization
• Vision understanding — read and reason about screenshots, diagrams, scanned documents, and UI mockups

Why Token Plan?

Pay-as-you-go pricing works — until it doesn't. Once you're running AI agents that chain multiple model calls, generating video with HappyHorse, or building voice-enabled applications, per-token billing becomes unpredictable and expensive.

The Token Plan solves this with a single monthly credit allocation that covers every model and modality on the platform. You get a predictable budget, dramatically lower effective cost per token, and the freedom to experiment across capabilities without worrying about a surprise bill.

Plans and Pricing

Three tiers designed for different usage levels, all with limited-time early-bird pricing:

How credit allocation works: Like Codex and other leading AI subscription products, the Token Plan uses a rolling reset model — not a flat monthly pool. Your credits refresh on two cycles:

• Every 5 hours: A fresh allocation becomes available, so you're never locked out for long after a heavy session.
• Every 7 days: Your weekly ceiling resets, ensuring consistent access throughout your billing period.

This means you always have credits available for your next work session — the plan is designed for sustained daily use, not a lump-sum budget you burn through in the first week.

Save More with Longer Billing Cycles

Commit quarterly or yearly to unlock additional savings on top of early-bird pricing:

The Standard tier at yearly billing works out to roughly $16.25/month — that's 10,000 credits per week (refreshing every 7 days) of multimodal AI for the price of a lunch.

Full Model Coverage — One Plan, Every Modality

Unlike single-model subscriptions, the Token Plan gives you access to the full spectrum of Model Studio capabilities with one API key:

Text & Reasoning

• Qwen3.8-Max-Preview — Latest flagship model, debuting exclusively on Token Plan. Excels at full-stack development, data analysis, and office workflows with built-in vision understanding.
• Qwen3.7-Max — Frontier-class agent model for coding, reasoning, and autonomous execution
• Qwen3.7-Plus — Cost-effective multimodal workhorse for high-volume production

Video Generation

• HappyHorse 1.1 — Latest video generation model for short-form content, animations, and brand ads. See our HappyHorse prompts guide for tips.

Audio & Speech

• Qwen-Audio-3.0-TTS-Plus — High-quality text-to-speech synthesis
• Fun-ASR — Automatic speech recognition and transcription

Image Generation

• Wan2.7-Image — Production-grade image generation for product photos, posters, and creative assets

Built-in Harness Tools

Your Token Plan credits also cover platform-native tools that your agents can call directly:

• Web search
• Text-to-image search
• Image-to-image search
• Web scraping
• Code interpreter

Built-in Skills

• Image generation (on-platform)
• Video generation (on-platform)

Unbeatable Value — 3x More Usage Plus Stacking Discounts

3x More Than Pay-As-You-Go

Dollar for dollar, the Token Plan delivers approximately 3x more usage compared to spending the same amount on pay-as-you-go API calls. The math is simple: you're buying credits in bulk at a substantial discount.

Off-Peak Discounts Stack

Here's what makes the Token Plan uniquely powerful: off-peak rates (22:00–08:00 Beijing Time) stack directly with your Token Plan credits. During off-peak hours, your credits stretch even further:

What this means in practice: A Standard plan user (10,000 credits/week) running Qwen3.7-Max exclusively during off-peak hours effectively gets 50,000 credits worth of usage per week — that's the equivalent of 18/month.

Combined Savings Example

Consider a developer on the Standard plan ($18/mo) who runs 70% of their workload during off-peak hours using Qwen3.7-Max:

• 30% at standard rate: 3,000 credits used normally per week
• 70% at 0.2x rate: 7,000 credits → effectively 35,000 credits of usage per week
• Total effective weekly usage: ~38,000 credits for $18/month

That's nearly 4x the Pro tier's weekly allocation at less than a third of the price — simply by timing your workloads smartly.

Flexible Plans — Three Tiers, Three Billing Cycles

Lite — For Individuals and Side Projects

700 credits per 5 hours / 2,500 credits per week is enough for regular AI-assisted coding, occasional image generation, and light experimentation. Perfect if you're:

• Using AI coding assistants for daily development
• Prototyping multimodal applications
• Running occasional video or image generation

Standard (Recommended) — For Active Builders

3,000 credits per 5 hours / 10,000 credits per week covers serious daily usage across multiple modalities. Built for developers who:

• Run AI agents throughout the workday
• Generate video content regularly
• Use multiple models across a single project
• Need headroom for burst usage

Pro — For Power Users and Small Teams

12,000 credits per 5 hours / 40,000 credits per week for high-volume production workloads. Designed for:

• Running multiple concurrent agents
• Heavy video and image generation pipelines
• Production-grade AI applications
• Teams sharing a single subscription

Upgrade Anytime

Start with Lite, move to Standard when you hit your stride, upgrade to Pro when you scale. No penalty for upgrading mid-cycle — your new allocation kicks in immediately. You don't need to commit big upfront.

Credit Reset Cadence and Concurrency

The Token Plan's rolling reset model keeps your credits fresh — here's what each tier provides:

Why Rolling Resets Matter

Unlike a flat monthly quota that encourages hoarding or binge-spending, the 5-hour + 7-day rolling reset ensures:

• You never run dry for long — even after an intense coding sprint, your next allocation is just hours away
• Predictable daily throughput — you can rely on a consistent credit flow for daily agent-driven workflows
• No end-of-month anxiety — credits don't deplete irreversibly; the system is built for sustained, active use

This is the same cadence used by Codex and other professional AI subscription products — it's purpose-built for developers who use AI tools every day, not once a month.

Concurrent Agents

Concurrent agent limits define how many AI tools or agents can use your API key simultaneously — critical if you're running multiple coding assistants or agent pipelines in parallel.

How Credits Work

Credits are the unified usage unit across all models on the Token Plan. Here's what you need to know:

• Different models consume credits at different rates — flagship models like Qwen3.8-Max-Preview cost more credits per call than lightweight models like Qwen3.7-Plus
• Credits reset on a rolling basis — your 5-hour allocation refreshes every 5 hours, and your weekly allocation refreshes every 7 days. Unused credits within a window do not carry over to the next.
• Off-peak multipliers reduce credit consumption — making your allocation stretch further during discount hours
• All modalities draw from the same pool — text, video, image, and audio all use your credit balance

Use Cases by Industry

The Token Plan's multimodal coverage makes it uniquely suited for cross-capability workflows:

Cross-Modality Workflow Example

A typical e-commerce AI workflow on the Standard plan:

1. Qwen3.8-Max-Preview generates product descriptions and marketing copy
2. Wan2.7-Image creates product lifestyle images from text prompts
3. HappyHorse 1.1 produces short video ads from product images
4. Qwen-Audio-3.0-TTS-Plus narrates the video with natural speech

All four steps draw from the same 40,000 credit pool — no separate subscriptions, no juggling API keys.

Getting Started

Step 1: Subscribe to a Plan

Visit the Model Studio Token Plan page and select your tier. Early-bird pricing is applied automatically during the launch period.

Step 2: Get Your Dedicated API Key

After subscribing, you'll receive a dedicated Token Plan API key in the format sk-sp-xxxxx. This key is separate from regular Model Studio API keys.

Step 3: Configure Your Tools

The Token Plan works with any OpenAI-compatible or Anthropic-compatible tool:

OpenAI-compatible endpoint:

https://coding.dashscope.aliyuncs.com/v1

Anthropic-compatible endpoint:

https://coding.dashscope.aliyuncs.com/apps/anthropic

Compatible tools include:

• Qwen Code
• Claude Code
• OpenClaw
• OpenCode
• Any OpenAI/Anthropic-compatible client

One API key works across all your tools simultaneously — no need to generate separate keys for each application.

Step 4: Build

That's it. Your credits are live, off-peak stacking is automatic, and every model on the platform is available through your single API key. Start building.

Key Takeaways

• Credits reset every 5 hours and every 7 days — a rolling refresh model (like Codex) designed for sustained daily use, not a flat monthly pool.
• One subscription covers every modality — text, video, image, audio, and built-in tools all draw from the same credit pool.
• 3x more usage compared to equivalent pay-as-you-go spend, before any additional discounts.
• Off-peak discounts stack — Qwen3.7-Max drops to 0.2x credit consumption during off-peak hours (22:00–08:00 Beijing Time), effectively giving you 5x the usage during that window.
• Early-bird pricing starts at $6/month for the Lite tier — limited time only.
• Works with your existing tools — OpenAI-compatible and Anthropic-compatible endpoints mean zero migration effort.
• Flexible scaling — upgrade from Lite to Standard to Pro anytime as your usage grows.

FAQ

What is the Model Studio Token Plan for Individual?

The Token Plan for Individual is a monthly credit-based subscription that gives you access to all Model Studio models — text, video, image, and audio — at a significantly lower effective cost than pay-as-you-go pricing. One subscription, one API key, every modality.

How much can I save compared to pay-as-you-go?

On average, Token Plan users get approximately 3x more usage for the same dollar amount. When combined with off-peak discounts (which stack with Token Plan credits), effective savings can reach 5x or more on supported models.

Do off-peak discounts really stack with the Token Plan?

Yes. During off-peak hours (22:00–08:00 Beijing Time), your credit consumption is reduced — Qwen3.7-Max uses only 0.2x credits and Qwen3.7-Plus uses only 0.4x credits. This is in addition to the baseline savings of the Token Plan itself. Learn more about off-peak rates.

What models are included?

All current Model Studio models: Qwen3.8-Max-Preview, Qwen3.7-Max, Qwen3.7-Plus (text/reasoning), HappyHorse 1.1 (video), Wan2.7-Image (image), Qwen-Audio-3.0-TTS-Plus, Fun-ASR (audio), plus built-in harness tools like web search, code interpreter, and web scraping.

Can I use my existing API key?

No. The Token Plan uses a dedicated API key format (sk-sp-xxxxx) that is separate from regular Model Studio API keys. You'll receive your dedicated key after subscribing.

Do unused credits roll over?

No. Credits refresh on a rolling basis — every 5 hours and every 7 days — and unused credits within a window do not carry forward to the next reset. The system is designed for consistent daily use, not accumulation. Off-peak stacking helps you maximize every credit during each window.

Can I upgrade my plan mid-cycle?

Yes. You can upgrade from Lite to Standard or Standard to Pro at any time. Your new credit allocation takes effect immediately.

Is the Token Plan for automated backend services?

No. The Token Plan is designed for interactive use with AI coding and agent tools — not for automated scripts or backend service integrations. Use standard pay-as-you-go API keys for programmatic production workloads.

Can I hold multiple Token Plan subscriptions?

No. One subscription per identity. You can upgrade your tier but cannot hold multiple active plans simultaneously.

What tools work with the Token Plan?

Any OpenAI-compatible or Anthropic-compatible tool works — including Qwen Code, Claude Code, OpenClaw, OpenCode, and any client that supports the OpenAI or Anthropic API format. One API key works across all tools simultaneously.

What happens when my credits run out?

Credits operate on two rolling windows: a 5-hour allocation and a 7-day allocation. If you exhaust either, service pauses until that window resets — the 5-hour window refreshes every 5 hours, and the weekly window resets every 7 days. If you consistently run out, consider upgrading to a higher tier for more headroom.

Is there a refund policy?

No refunds are available for Token Plan subscriptions. We recommend starting with the Lite tier to evaluate fit before committing to a higher plan or longer billing cycle.

Get Started Today

The Token Plan is the most cost-effective way to access every AI capability on Model Studio — text, video, image, and audio — through a single subscription. Early-bird pricing won't last forever.

1. Get your promo coupon — Claim your coupon here for extra savings on top of early-bird pricing
2. Choose your tier — Subscribe to Lite (18/mo), or Pro ($68/mo)
3. Get your sk-sp-xxxxx API key from the console
4. Connect your tools — point any OpenAI/Anthropic-compatible client to the endpoint
5. Build more, spend less — your credits are live and off-peak stacking is automatic

For more on the AI tools and models available on Model Studio, explore our guides on OpenAI API alternatives, AI agents, and HappyHorse video generation.