Answer: The latest stable PyTorch version is 2.15.0, and the listed CUDA compute platforms are CUDA 13.2, CUDA 13.4, and CUDA 12.8. Status: Check failed Entry: version/pytorch-cuda Revision: 1 Digest: sha256:1f82592ea61ef331 (changes whenever the content changes) Timeline: Created: 2026-11-12T11:08:22Z Last changed: 2026-11-12T11:08:22Z Last verified: 2026-11-12T11:23:53Z Last attempt: 2026-11-12T11:34:04Z (failed) Next check: 2026-11-12T12:34:04Z Check interval: Daily Expected change: No fixed schedule Data: Field | Value | Source | As of PyTorch version | 2.15.0 | [1] | 2026-11-12 CUDA version | CUDA 13.2 | [1] | 2026-11-12 Sources: [1] PyTorch Foundation. "Get Started". https://pytorch.org/get-started/locally/ Role: primary. Accessed: 2026-11-12T11:23:54Z Retrieved via: DeepSeek Fetch API (Markdown conversion; the original server response was not captured) Quote: "PyTorch Build Stable (2.15.0) Preview (Nightly)" (Supports: PyTorch version) Quote: "Compute Platform CUDA 13.2 CUDA 13.4 CUDA 12.8 ROCm 10.0 CPU" (Supports: CUDA version) Snapshot SHA-256: 9c7fdccd005a607d08ccb8c9192ce3ba1a70df76a21e7eded78a9c98641b4f59 (captured 2026-11-12T11:23:54Z) Verification: Method: Automated extraction from [1] with deepseek-flash (JSON output, thinking disabled). Every value was checked by code to appear verbatim in a quoted passage of [1]. [1] was retrieved through the DeepSeek Fetch API because a direct request did not return usable content. Cross-check: not-available at 2026-11-12T11:23:53Z. No independent second source is configured for this entry Last successful check: 2026-11-12T11:23:53Z Last attempt: 2026-11-12T11:34:04Z (failed) Notes and limitations: Extracted automatically from a single official source. Change log: Time | Event | Change | Source 2026-11-12T11:34:04Z | check-failed | Extraction did not pass verification: answer contains "12.6", which does not appear in any quoted passage | — 2026-11-12T11:08:22Z | added | Entry created | — Changes for this entry as a feed: https://latestn.com/version/pytorch-cuda/latest Challenge this entry: Open challenges: None How: If a value here is wrong or out of date, send what is wrong and a link to evidence. Challenges are queued and checked against the evidence; they do not change the entry by themselves. If the entry changes as a result, the change log says so. Form: https://latestn.com/dispute/version/pytorch-cuda POST: https://latestn.com/dispute/version/pytorch-cuda with claim, evidence_url, and optionally field and contact (form-encoded or JSON) GET: https://latestn.com/dispute/version/pytorch-cuda?claim={what is wrong}&evidence_url={link} How to cite: latestN. (2026). What is the latest PyTorch version, and which CUDA versions does it support? Revision 1, last verified 2026-11-12T11:23:53Z. https://latestn.com/version/pytorch-cuda (accessed 2026-11-12T12:52:51Z). @misc{latestN_version_pytorch_cuda_r1, title = {What is the latest PyTorch version, and which CUDA versions does it support?}, author = {{latestN}}, year = {2026}, howpublished = {\url{https://latestn.com/version/pytorch-cuda}}, note = {Revision 1; digest sha256:1f82592ea61ef331; last verified 2026-11-12T11:23:53Z; accessed 2026-11-12T12:52:51Z} } URL: https://latestn.com/version/pytorch-cuda