Alibaba has placed Qwen3.8-Max-Preview on its Token Plan and Qoder developer surfaces as a hosted preview, while the underlying Qwen3.8 open-weight release with a stated 2.4 trillion parameters remains announced but not yet public. According to the LinkLoot blog's reading of Qwen's official X post from July 19, 2026, q For developers, the practical takeaway is that Qwen3.8-Max-Preview is reachable today through Alibaba's Token Plan API (ap-southeast-1.maas.aliyuncs.com compatible-mode endpoint), Qoder and QoderWork IDE integrations, and the Qianwen PC client, but the open-weight drop, full model card, and independent benchmarks are s

A KuCoin flash dated July 19, 2026 confirms the same Aliyun Qwen announcement: the Qwen3.8 model with 2.4 trillion parameters is being prepared for open-source release, and a Qwen3.8-Max-Preview is already available on Alibaba Token Plan as well as Qoder and QoderWork for developer testing. The item corroborates that Alibaba Token Plan currently hosts a preview of the Qwen3.8-Max model, giving developers an early-access channel for the 2.4T-parameter release. As with the underlying wire, it provides no further API, pricing, or rollout-date details, but it reinforces the immediacy of preview availability o

Alibaba Cloud's Qwen team announced on July 19, 2026, an upcoming open-source release of Qwen3.8, an ultra-large-scale model featuring 2.4 trillion parameters, with the team stating its overall capabilities rival cutting-edge AI systems and self-assessed as ranking just behind "Fable 5." For developers using Alibaba To The news confirms a direct, time-sensitive addition to the Alibaba Token Plan model lineup: a preview of the 2.4T-parameter Qwen3.8-Max is accessible to developers on the platform now, even though the full open-source release timeline and API specifics (endpoints, pricing, context length) were not disclosed in the repo