When was Qwen 2.5 released?
Qwen announced the Qwen2.5 model family on September 19, 2024. The announcement introduced Qwen2.5 alongside the specialized Qwen2.5-Coder and Qwen2.5-Math models. (Qwen’s official announcement)
Release date vs. technical report dates
The Qwen2.5 Technical Report was submitted to arXiv on December 19, 2024, and revised on January 3, 2025. Those are the report’s submission and revision dates—not the Qwen2.5 family’s release date. (arXiv record)
What the release date means for choosing a Qwen API
The September 2024 announcement identifies when the model family was introduced. It does not establish which hosted Qwen model is currently available, what its current capabilities or limits are, or whether it fits your production workload.
Qwen’s announcement historically mentioned Qwen-Plus and Qwen-Turbo APIs through Model Studio. (Qwen’s announcement) For a present-day starting point, Alibaba Cloud’s Model Studio directory lists newer options such as qwen3.7-plus, qwen3.8-max, and qwen3.8-flash. (Model Studio model directory) Its text-generation guidance describes Plus as a balance of capability and cost, Max for higher reasoning and understanding needs, and Flash as a cost-oriented option whose results should be checked against your requirements. (Model Studio text-generation guidance) Treat these as starting points, then evaluate the current model against your workload.
Before deploying, check the current first-party documentation for the specific model’s identifier, features, and limits. Model names and availability can change; the release date alone cannot settle those choices. (Model Studio model directory)
Sources
Qwen2.5: A Party of Foundation Models!