Qwen 3.8 Flash Next TensorFold runtime artifacts

Pinned TensorFold 0.3.6.3 runtime for Thor SM 11.0, with the tested 4-stream, 312K-context profile and its matching precompiled CUDA, Triton and NVIDIA warm cache.

These archives contain the runtime and warm cache. Model weights are downloaded separately from Vontra/Qwen3.8-Flash-Next-MLX-4bit-MTP or its ModelScope mirror.

The runtime image ID is sha256:58bbc0c715544b26a53409edac564e108f6d58ea47024110e3f948628e92ef68. See artifact-manifest.json for exact sizes, SHA256 checksums and the warm-cache fingerprint. The archives preserve the published runtime 0.1.7 bytes; LPK 0.1.13 selects either Hugging Face or ModelScope for all deployment downloads.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support