Issues

• [CLI] smollm3-3b chat model failing with simple prompt

  microsoft/foundry-local#1243 · Filter by author OhShenbanjo opened on Oct 26, 2026

• [CLI] Phi-4 Mini CPU variant unexpectedly requires CUDAExecutionProvider

  microsoft/foundry-local#1242 · Filter by author kpetersonNW opened on Oct 26, 2026

• [CLI] "error: The Foundry Local runtime reported an internal error." for openvino npu and gpu models

  microsoft/foundry-local#1241 · Filter by author iSatishYadav opened on Oct 26, 2026

• Foundry Local 0.12.1: WinError 5 when loading CPU runtime

  microsoft/foundry-local#1240 · Filter by author KentHDBlue opened on Oct 26, 2026

• [CLI] NPU model loading fails with OpenVINO EP API error

  microsoft/foundry-local#1237 · Filter by author kemlik opened on Oct 25, 2026
  1

• [Feature Request] Request for more NPU-optimized models (QNN) on Snapdragon X Elite

  enhancement
  microsoft/foundry-local#1236 · Filter by author TK-YKHM opened on Oct 24, 2026

• Consume patched ORT and CUDA artifacts for Windows telemetry shutdown hang

  microsoft/foundry-local#1233 · Filter by author bmehta001 opened on Oct 23, 2026

• [Feature Request] Add Google Gemma 4 E4B Support to Foundry Local

  enhancement
  microsoft/foundry-local#1232 · Filter by author goxia opened on Oct 23, 2026

• phi-3.5-mini-instruct-qnn-npu:3 and phi-3-mini-128k-instruct-qnn-npu:4 failing on Qualcomm X2Elite

  microsoft/foundry-local#1227 · Filter by author pathakpragnya03 opened on Oct 21, 2026
  1

• [Bug] Windows native runtime ignores HTTP_PROXY/HTTPS_PROXY for SDK networking

  bug
  microsoft/foundry-local#1221 · Filter by author jiec-msft opened on Oct 20, 2026

• Restore experimental Java SDK CI after blocked Maven bootstrap downloads

  microsoft/foundry-local#1219 · Filter by author baijumeswani opened on Oct 20, 2026

• [CLI] NPU models fail on Snapdragon X1P42100: "EPContext node ... not compatible with any execution provider"

  microsoft/foundry-local#1215 · Filter by author KaranMegha opened on Oct 17, 2026

• [Bug] CUDA EP fails to load under Smart App Control: onnxruntime-genai-cuda.dll in the CUDA EP bundle is not Authenticode-signed (error 4551)

  microsoft/foundry-local#1210 · Filter by author Ooooze opened on Oct 16, 2026

• Update docs and samples to remove the WinML package and the obsolete Betalgo-based APIs

  microsoft/foundry-local#1208 · Filter by author petertorr opened on Oct 15, 2026
  2

• [Bug] New ChatSession API is massive regression (perf and usefulness) compared to old GetChatClientAsync() API.

  bug
  microsoft/foundry-local#1207 · Filter by author petertorr opened on Oct 15, 2026
  2

• [Bug] Model smollm3-3b always reports an error "Unknown method: replace at row 23, column 48:" for chat completions

  bug
  microsoft/foundry-local#1200 · Filter by author petertorr opened on Oct 15, 2026

• [Bug] Chat completion fails with "an illegal memory access was encountered"

  bug
  microsoft/foundry-local#1199 · Filter by author petertorr opened on Oct 15, 2026
  

• [Bug] Several models will fail if you try to load a generic-gpu variant if they also have a cuda variant

  bug
  microsoft/foundry-local#1198 · Filter by author petertorr opened on Oct 14, 2026
  1

• [Bug] Model gemma-4-e2b-it won't load on NVIDIA GPU

  bug
  microsoft/foundry-local#1197 · Filter by author petertorr opened on Oct 14, 2026

• [Bug] Most of the models that claim to support tools-calling do not, in fact, support tools-calling

  bug
  microsoft/foundry-local#1195 · Filter by author petertorr opened on Oct 14, 2026

• [Bug] Setting ChatClient.Settings.ToolChoice = Required never results in a correct tools call

  bug
  microsoft/foundry-local#1194 · Filter by author petertorr opened on Oct 14, 2026

• [Bug] The CompleteChatAsync API takes a CancellationToken but ignores it.

  bug
  microsoft/foundry-local#1193 · Filter by author petertorr opened on Oct 14, 2026
  2

  

• [Bug] onnxruntime_providers_cuda.dll throws unhandled C++ exception while completing chat request that terminates the process

  bug
  microsoft/foundry-local#1191 · Filter by author petertorr opened on Oct 14, 2026

• [CLI] <brief issue title>

  microsoft/foundry-local#1186 · Filter by author kukimolina opened on Oct 9, 2026

• The QWEN models don't support tool calling for QNN provider, but do for CUDA

  microsoft/foundry-local#1183 · Filter by author petertorr opened on Oct 8, 2026