Token Plan 团队版 vs Standard API Key

Sources:

- https://platform.qianwenai.com/pricing/token-plan

- https://platform.qianwenai.com/home/billing/subscription/token-plan

Updated: 2026-05-01

Two Key Types

千问云 has two mutually exclusive authentication systems. Mixing them produces hard-to-diagnose errors.

| Dimension | Standard Key (Pay-as-you-go) | Token Plan 团队版 | |-----------|------------------------------|-------------------| | Key format | sk-xxxxx | sk-sp-xxxxx | | OpenAI-compatible URL | https://dashscope.aliyuncs.com/compatible-mode/v1 | https://token-plan.cn-beijing.maas.aliyuncs.com/compatible-mode/v1 | | Anthropic-compatible URL | N/A | https://token-plan.cn-beijing.maas.aliyuncs.com/apps/anthropic | | Native DashScope URL | https://dashscope.aliyuncs.com/api/v1 | Not supported on text endpoints; image generation uses dedicated https://token-plan.cn-beijing.maas.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation | | Auth header | Authorization: Bearer | Authorization: Bearer (NOT x-api-key) | | Supported text models | Full catalog (100+) | 4 text LLMs (see below) | | Supported image models | Full catalog | 4 image models via Skill/extension only (see below) | | Video / TTS / ASR / Embedding | Available | Not supported | | Usage scope | Any API call (scripts, apps, tools) | Interactive AI tools only (Cursor, Claude Code, Qwen Code, OpenClaw, OpenCode, Codex, Kilo Code/CLI, Hermes Agent, etc.) | | Billing | Per-token consumption (CNY) | Credits: monthly seat allowance + shared usage packages | | Quota exhaustion | Continues (pay more or use prepaid balance) | Hard fail — service paused until next cycle or shared package purchased |

[!WARNING]

Forbidden uses for sk-sp- Token Plan keys: automation scripts, application backends, batch jobs,

API testing tools (Postman/Insomnia), workflow platforms (Dify, n8n, Coze). Violations may trigger

subscription suspension or API Key revocation.

Supported Models

Text Models (4 total)

| Model | Context Window | OpenAI-compat | Anthropic-compat | Notes | |------------------|---------------:|:-------------:|:----------------:|--------------------------------| | qwen3.6-plus | 1M | ✅ | ✅ | Flagship; multimodal text+image input; thinking mode (thinkingFormat: qwen); built-in tools (web search, code interpreter, web fetch, image search) via Responses API | | glm-5 | 198K | ✅ | ✅ | Thinking mode supported | | MiniMax-M2.5 | 192K | ✅ | ✅ | budgetTokens + output ≤ 32768 | | deepseek-v3.2 | 128K | ✅ | ❌ OpenAI only | Use compatible-mode/v1 only |

Image Generation Models (4 total)

[!IMPORTANT]

Image generation models cannot be invoked through the text Base URL. They use a dedicated

multimodal-generation endpoint and must be wired up via each tool's Skill / Slash Command / Agent

mechanism (NOT through the standard model selector).

| Model | Notes | |----------------------|--------------------------------------------------------------------| | qwen-image-2.0 | Default; general-purpose; strong Chinese text rendering | | qwen-image-2.0-pro | Higher quality, slightly slower | | wan2.7-image | Multi-style; returns 4 images by default | | wan2.7-image-pro | Supports 4K (additional sizes: 2048×2048, 1440×2560, 2560×1440) |

Available sizes: 10241024 (default), 7201280, 1280*720. wan2.7-image-pro adds 4K options above.

Endpoint: POST https://token-plan.cn-beijing.maas.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation

Credits Billing Mechanism

- Unit: Credits. Single-call cost depends on model, token usage, thinking mode, and tool calls.

• Tiers: 标准坐席 (¥198/seat/month, 25,000 Credits) · 高级坐席 (¥698/seat/month, 100,000 Credits) · 尊享坐席 (¥1,398/seat/month, 250,000 Credits).
• Shared usage package: ¥5,000 per package, 625,000 Credits, 1-month validity, expires unused.
• Deduction order: seat monthly quota → shared package (nearest-expiry first) → service paused.
• Reset: seat quotas reset monthly; unused credits do not roll over.

  Example (qwen3.6-plus single request): 8,349 input + 40,794 cached + 573 output ≈ 3.18 Credits.

  Key Type × Endpoint — Expected Behavior

  | Key Type | Base URL | Result |

|----------|----------|--------| | sk- | dashscope.aliyuncs.com/... | OK | | sk- | token-plan.cn-beijing.maas.aliyuncs.com/... | Incorrect API key provided — standard key rejected on Token Plan endpoint | | sk-sp- | dashscope.aliyuncs.com/... | InvalidApiKey: Invalid API-key provided — Token Plan key rejected on standard endpoint | | sk-sp- | token-plan.cn-beijing.maas.aliyuncs.com/... (via tool) | OK | | sk-sp- | token-plan.cn-beijing.maas.aliyuncs.com/... (via raw script/curl outside an AI tool) | Policy violation — may trigger suspension |

Common Errors

| Error | Cause | Resolution | |--------------------------------------------------|------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------| | InvalidApiKey: No API-key provided | Key not configured, or tool used x-api-key header | Set key; switch to Authorization: Bearer | | InvalidApiKey: Invalid API-key provided | Standard sk- key used, subscription expired, key copied with whitespace | Use sk-sp- Token Plan key; verify subscription status; reset key in console | | model 'xxx' not found or not supported | Model name typo / wrong case; model not supported on chosen protocol | Match model ID exactly; deepseek-v3.2 only via OpenAI-compat | | invalid access token or token expired | Wrong Base URL (e.g. used another plan's endpoint) | Use Token Plan Base URL listed above | | Incorrect API key provided | Used dashscope.aliyuncs.com Base URL with Token Plan key | Switch to token-plan.cn-beijing.maas.aliyuncs.com endpoint | | Range of input length should be [1, xxx] | Input + history exceeds context window | Start a new session, compact context, or switch to a larger-context model | | Connection error | Base URL typo or network issue | Verify Base URL spelling and connectivity | | API rate limit reached | Routed to shared API quota due to mis-config; or seat / shared package exhausted | Verify provider config; check Token Plan console for usage; reset key as last resort |

Impact on QianWen-AI/qianwen-ai Scripts

All execution scripts (qianwen-text, qianwen-vision, qianwen-image-generation, qianwen-video-generation, qianwen-audio-tts) call DashScope directly via urllib.request and are not recognized as AI tools. They detect sk-sp- keys and print a warning to stderr.

| Skill | API Type | Works with sk-sp- Token Plan key? | Reason | |-----------------------------|---------------|:-----------------------------------:|-------------------------------------------------------------------------------------| | qianwen-text | OpenAI-compat | ❌ | Standard endpoint rejects sk-sp-; scripts are not AI tools | | qianwen-vision | OpenAI-compat | ❌ | Same; vision models not in Token Plan list | | qianwen-image-generation | Native | ❌ | Standard dashscope.aliyuncs.com not accepted by Token Plan; image models on Token Plan need separate endpoint via tool Skill mechanism | | qianwen-video-generation | Native | ❌ | Video models unavailable in Token Plan | | qianwen-audio-tts | Native | ❌ | TTS models unavailable in Token Plan |

Action: Set a standard sk- key in DASHSCOPE_API_KEY when using these execution skills. The Token Plan sk-sp- key is for the AI tool itself (Cursor, Claude Code, etc.), not for backend scripts.

Cost Risk Scenarios

1. sk-sp- key in scripts: 401/Incorrect API key provided, no charges, but confusing. 2. sk- key when user expects Token Plan coverage: Calls succeed but incur pay-as-you-go charges. Cannot detect programmatically — documentation must clarify. 3. QWEN_BASE_URL set to Token Plan endpoint with sk- key: Authentication fails. 4. Token Plan Credits exhausted: Hard fail, no fallback to pay-as-you-go.

Console & Billing

| Resource | URL | |---------------------------|--------------------------------------------------------------------| | Token Plan Subscription | https://platform.qianwenai.com/home/billing/subscription/token-plan | | Token Plan Pricing | https://platform.qianwenai.com/docs/token-plan/overview#%E5%A5%97%E9%A4%90%E4%B8%8E%E5%AE%9A%E4%BB%B7 | | Pay-as-you-go Billing | https://platform.qianwenai.com/home/billing/pay-as-you-go | | Usage Analytics (PAYG) | https://platform.qianwenai.com/home/analytics |

[!NOTE]

Usage queries: Token Plan seat & shared-package Credits balance are currently only viewable in

the Token Plan console (Subscription page →

Token Plan tab). The qianwen CLI does not yet support sk-sp- Token Plan team-edition keys; CLI

commands (qianwen usage summary, etc.) only work for standard sk- keys.

Coexistence

Both key types can be held simultaneously by the same user:

• sk-sp- Token Plan key → configured in the AI tool (Cursor, Claude Code, OpenClaw, ...)
• sk- standard key → set in DASHSCOPE_API_KEY for QianWen-AI/qianwen-ai execution scripts

  These are independent; configuring one does not affect the other.