Alibaba releases open Qwen3.8 with 2.4T total, 95B active parameters
Alibaba released its 2.4-trillion-parameter Qwen3.8 model as open weights, giving operators a text-only checkpoint with about 95 billion parameters active per token.
By RuntimeWire Staff · Published
Primary source: Hugging Face
Why it matters
Qwen3.8 gives AI teams access to the weights for a Qwen-Max-class model, but its scale requires substantial serving infrastructure. The downloadable checkpoint is text-only, always uses thinking mode and activates about 95 billion parameters per token.
Developers and infrastructure teams can now download the model weights for Qwen3.8, the latest release from Alibaba's large-model program. Alibaba published the Qwen3.8-2.4T-A95B weights on August 12. Qwen calls it the first Qwen-Max-class model released openly in its Qwen3.8 announcement.
The repository provides Qwen3.8-2.4T-A95B as model weights and configuration files in Hugging Face Transformers format, with compatibility for vLLM, SGLang and TokenSpeed. Qwen separately offers Qwen3.8-Max through Qwen Cloud. According to Qwen's repository and cloud documentation, that hosted version supports vision input, non-thinking mode and built-in tools.
What Qwen released
Qwen's repository describes Qwen3.8-2.4T-A95B as a 2.4-trillion-parameter sparse mixture-of-experts model with about 95 billion parameters activated per token. Its 92 layers contain 512 experts, with 10 routed experts and one shared expert active at a time. That design increases the model's stored capacity without using all 2.4 trillion parameters for every inference step.
The repository says the open checkpoint is text-only, requires thinking mode and supports a native 262,144-token context that can be extended to 1,010,000 tokens. Its reasoning_effort setting supports xhigh by default, medium and low. A separate preserve_thinking control retains reasoning context from earlier messages.
The model card lists support for Hugging Face Transformers, SGLang, vLLM and TokenSpeed.
This is far removed from the local-development case for smaller Qwen releases. RuntimeWire reported in June that Qwen 3.6 27B had become practical for local coding work. Qwen3.8's large parameter count creates a substantial storage requirement before an operator accounts for serving memory, compute and inference throughput.
The open checkpoint is narrower than Qwen3.8-Max
The downloadable checkpoint is text-only and always operates in thinking mode. It does not accept image or video inputs, and developers cannot disable its reasoning mode through the documented interface.
According to Qwen's model repository, Qwen3.8-Max adds vision input, non-thinking mode, built-in tools and a default 1-million-token context window. The same repository directs users seeking managed, scalable inference without infrastructure maintenance to Qwen Cloud.
The two releases give developers a downloadable text model and a hosted Max product with additional input modes and tools.
The repository links to a model-specific license
The repository links to a Qwen3.8-specific license file. The supplied materials do not include the license text, so its commercial-use conditions cannot be summarized here.
What the published benchmarks establish
The supplied materials do not establish independent benchmark results for the downloadable Qwen3.8-2.4T-A95B checkpoint, while Qwen's model card reports company-run comparisons for the hosted Qwen3.8-Max version. Those comparisons cover coding, agent and general-capability tests under varying harnesses, run counts and evaluation conditions. Performance and operating costs for third-party deployments of the downloadable checkpoint therefore remain unestablished by the supplied evidence.