← All models

Qwen3.8-27B

Published Publisher Alibabaqwen3.8-27b

Overview

Qwen3.8-27B is an open-weight, 27-billion-parameter dense vision-language model in Alibaba’s Qwen3.8 series. It is intended for coding, office productivity, research, and multi-step agentic workflows. The weights are released under Apache 2.0. (alibabacloud.com)

Capabilities and limits

  • Accepts text, images, and video; produces text. Alibaba Cloud Model Studio lists function calling, structured output, web search, prefix completion, and context caching as supported. (alibabacloud.com)
  • Supports thinking and non-thinking modes. The open-weight model has a native context length of 262,144 tokens, extendable to 1 million tokens using YaRN. The model card cautions that static YaRN can affect performance on shorter texts. (huggingface.co)
  • Model Studio lists a 1,000,000-token context window, maximum input of 991,808 tokens, or 983,616 in thinking mode, and maximum output of 131,072 tokens. Supported length can vary with API parameters. Batch inference and model fine-tuning are listed as unsupported for this hosted service; this does not preclude fine-tuning the open weights with other tools. (alibabacloud.com)
  • Outputs can be incorrect and should be checked for consequential use. Local deployment requires suitable computing resources; Alibaba notes that quantization can enable laptop use. (alibabacloud.com)

Release

Released on August 14, 2026, according to the official Qwen repository. The official model card provides deployment and configuration details. (github.com)

Model Studio API pricing

Prices below are official list prices per 1 million tokens, checked November 16, 2026. They exclude limited-time promotions. Standard prices apply to both thinking and non-thinking modes for the documented 0–1 million input-token range; output billing includes reasoning and answer tokens. Regional and account-site prices are distinct published offers, not currency conversions. (help.aliyun.com)

Official site and region Currency Standard input Standard output Implicit-cache input Explicit-cache creation Explicit-cache read
Mainland site, China (Beijing) CNY 3 12 0.6 3.75 0.3
Mainland site, Singapore, international deployment CNY 3.646 21.875 0.729 4.557 0.365
International site, China (Beijing) USD 0.424 1.696 0.085 0.53 0.042
International site, Singapore, international deployment USD 0.50 3.00 0.10 0.625 0.05

Sources: Alibaba Cloud mainland model page and international model page. (help.aliyun.com)