Model prices / Alibaba / Qwen3.5 27B
Qwen3.5 27BOpen weights
Alibaba · released 2026-02-23
The official API price of Qwen3.5 27B is $0.3 per 1M input tokens and $2.4 per 1M output tokens (as of 2026-11-12). Context window: 262K tokens, up to 65.5K output tokens per request. 20 providers sell it. The lowest is AIHubMix: $0.0846 input, $0.6768 output.
Price by providerUSD per 1M tokens
| Provider | Input | Output | Cached | Notes | |
|---|---|---|---|---|---|
| AlibabaOfficialqwen3.5-27b | $0.3 | $2.4 | — | Official | Official docs |
 AIHubMixLowestTiered pricingqwen3.5-27b | $0.0846 | $0.6768 | — | LowestTiered pricing | Official docs |
 EmpirioLabs AITiered pricingqwen3-5-27b | $0.086 | $0.688 | $0.086 | Tiered pricing | Official docs |
MMerge Gatewayqwen/ | $0.086 | $0.688 | $0.0172 | Official docs | |
OOrcaRouterTiered pricingqwen/ | $0.086 | $0.688 | — | Tiered pricing | Official docs |
 SiliconFlow (China)Tiered pricingQwen/ | $0.089668 | $0.717348 | — | Tiered pricing | Official docs |
KKilo Gatewayqwen/ | $0.195 | $1.56 | — | Official docs | |
 ChutesQwen/ | $0.2 | $1.75 | $0.02 | Official docs | |
NNanoGPTqwen/ | $0.2 | $1.75 | $0.02 | Official docs | |
NNanoGPTqwen/ | $0.2 | $1.75 | $0.02 | Official docs | |
| SiliconFlowQwen/ | $0.25 | $2 | — | Official docs | |
OOfoxbailian/ | $0.29 | $2.05 | $0.29 | Official docs |
10 more listings
| Provider | Input | Output | Cached | Notes | |
|---|---|---|---|---|---|
OOfoxqwen/ | $0.29 | $2.05 | $0.29 | Official docs | |
 Hugging FaceQwen/ | $0.3 | $2.4 | — | Official docs | |
JJalapeno CloudQwen3.5-27B | $0.3 | $2.4 | — | Official docs | |
 DevPass (LLM Gateway)qwen3.5-27b | $0.3 | $2.4 | — | Official docs | |
LLLM Gatewayalibaba/ | $0.3 | $2.4 | — | Official docs | |
MMixlayerqwen/ | $0.3 | $2.4 | — | Official docs | |
 Novita AIqwen/ | $0.3 | $2.4 | — | Official docs | |
 Deep InfraQwen/ | $0.26 | $2.6 | — | Official docs | |
 OpenRouterqwen/ | $0.26 | $2.6 | — | Official docs | |
RRequestyqwen3.5-27b | $0.26 | $2.6 | — | Official docs |
Providers are those listed by models.dev. This site has not vetted them one by one; each provider's own page has the binding price.
About
Qwen vision-language model for visual reasoning, documents, and agent tasks
Description from models.dev.
Specs
- Type
- Text
- Context
- 262K
- Max output
- 65.5K
- Input
- Text, Image, Video
- Output
- Text
- Reasoning
- Yes
- Tool calling
- Yes
- Structured output
- Yes
- Knowledge cutoff
- 2025-04
- Open weights
- Yes
- Released
- 2026-02-23
Source
Prices and specs come from models.dev (MIT licence), updated 2026-11-12. Each provider's name links to its official docs. A provider priced at 0 usually includes the model in a subscription; it is not necessarily a free API.