Alibaba Cloud DashScope (Qwen) API API Pricing & Free Tier 2026
Verified May 14, 2026DashScope API is Alibaba Cloud's LLM platform — Qwen LLM, Qwen-VL multimodal, text embeddings, TTS, and other Alibaba self-developed models.
Free tier
Each Qwen tier ships 1M free tokens on signup, valid 180 days — usable across Qwen-Turbo, Plus, and Max independently.
Qwen-Turbo / Plus / Max each include 1M free tokens, valid 180 days from signup. Quotas do not stack across models.
Per-model pricing
| Model | Context | Input $/1M | Output $/1M | Notes |
|---|---|---|---|---|
| Qwen-Max | 128K | $1.40 | $5.60 | Flagship — approaches GPT-4 on Chinese benchmarks |
| Qwen-Plus | 128K | $0.40 | $1.20 | Mid-tier balance of cost and capability |
| Qwen-Turbo | 8K | $0.04 | $0.12 | Low cost; suited for high-frequency simple tasks |
vs Global rivals
- Qwen-Max vs GPT-4o (openai)Qwen-Max approaches GPT-4o on Chinese tasks at ~28% of its input price.
- Qwen-Max vs Claude 3.5 Sonnet (anthropic)Qwen-Max beats Claude 3.5 Sonnet on Chinese; trails on English and code reasoning.
- Qwen-Max vs Gemini 1.5 Pro (google)Comparable pricing; Qwen-Max edges Chinese, Gemini 1.5 Pro edges multimodal.
Rate limits & access
Rate limit: Per-model RPM/TPM; request higher in Aliyun console.
Rate limits scale per Aliyun account verification tier. Mainland-China region only by default; international endpoints require separate request.
FAQ
- Is Qwen stronger than GPT-4 on Chinese benchmarks?
- On some Chinese-specific benchmarks Qwen-Max slightly beats GPT-4; overall on par. Qwen is more sensitive to Chinese cultural context.
- Can the open-source version run locally?
- Yes. Download Qwen2.5 series from HuggingFace; ollama / vLLM both support. Watch license terms for commercial use.
Prices in USD. Some providers (DashScope, Doubao) bill in CNY; figures shown at ~7.2 CNY/USD reference rate. Verify on the provider's official page before procurement.