Issues

• [Bug] Vulkan ErrorOutOfDeviceMemory on Windows with AMD Radeon HD 8570 despite --gpu none

  docker/model-runner#1073 · Filter by author PORTRAITART1 opened on Sep 27, 2026
  1

• Docker Model Runner (llama.cpp backend) returns HTTP 500 "does not match the expected peg-native format" when a single chat turn requires 2+ sequential tool/function calls

  docker/model-runner#1063 · Filter by author vincecase opened on Sep 18, 2026
  1

• OrcaRouter provider support for the Docker Model Runner gateway

  docker/model-runner#1059 · Filter by author kuswardhanietidims-svg opened on Sep 9, 2026
  1

• Using CPU inference there is no way to control how much ram of the host is used

  docker/model-runner#1058 · Filter by author sbe-arg opened on Sep 8, 2026
  2

• Docker Model Runner: docker model bench leaves the loaded llama.cpp model returning only ? tokens (CUDA, GTX 1650)

  docker/model-runner#1056 · Filter by author JurminOnem opened on Sep 4, 2026
  1

• Docker Model Runner on Windows fails to provision com.docker.nv-gpu-info.exe, breaking llama.cpp CUDA backend after Docker Desktop 4.82.0

  docker/model-runner#1054 · Filter by author tkdesign opened on Sep 1, 2026
  1

• SSRF fix in PR #1038 blocks all private/loopback/link-local realm addresses, breaking internal/corporate registries

  docker/model-runner#1050 · Filter by author okhowang opened on Aug 24, 2026
  3

• /api/show returns no capabilities and empty details fields

  docker/model-runner#1046 · Filter by author yohasebe opened on Aug 22, 2026
  1

• ai/functiongemma tool calls are not parsed into tool_calls

  docker/model-runner#1045 · Filter by author yohasebe opened on Aug 22, 2026
  1

• diffusers: generation continues after client disconnect and blocks the event loop — no cancellation

  docker/model-runner#1041 · Filter by author maumorales opened on Aug 15, 2026
  1

• Gemma 4 on llama: Reasoning budget 0 does not work

  docker/model-runner#1039 · Filter by author MarcOlivierRodrigue opened on Aug 11, 2026
  1

• compose up: --backend flag is accepted but never applied

  docker/model-runner#1036 · Filter by author VedantMadane opened on Aug 10, 2026
  1

• Compose: allowlisted backend environment variables for inference process

  docker/model-runner#1035 · Filter by author VedantMadane opened on Aug 10, 2026
  1

• Disable thinking if reasoning_effort is none.

  docker/model-runner#1028 · Filter by author MarcOlivierRodrigue opened on Jul 27, 2026
  1

• configure --context-size applied nondeterministically; packaged 128K default loaded instead → kernel panic on 24GB Mac (3×)

  docker/model-runner#1025 · Filter by author tomaDev opened on Jul 24, 2026
  1

• Pascal GPUs (sm_61/sm_62) still fall back to CPU with docker model run despite #986

  docker/model-runner#1024 · Filter by author FlorentPoinsaut opened on Jul 23, 2026
  1

• vLLM support broken on Windows with WSL2

  docker/model-runner#1018 · Filter by author VIEWVIEWVIEW opened on Jul 18, 2026
  3

• Disk Usage Duplication

  docker/model-runner#1015 · Filter by author HoracioImagen opened on Jul 14, 2026
  3

• automatic unloading mechanism in Docker Model, similar to the one in Ollama

  docker/model-runner#1014 · Filter by author brunocarvalhodearaujo opened on Jul 14, 2026
  1

• More apple metal woes

  docker/model-runner#1009 · Filter by author Lewiscowles1986 opened on Jul 12, 2026
  4

• Available unified memory in llama.cpp

  good first issue
  docker/model-runner#994 · Filter by author alex-mashin opened on Jul 3, 2026
  1

• Docker pull fails for ai/llama3.2 with TLS handshake timeout (500 Internal Server Error)

  docker/model-runner#989 · Filter by author MrMiningIR opened on Jun 28, 2026
  1

• CUDA version requirement is too new

  docker/model-runner#978 · Filter by author hao10 opened on Jun 19, 2026
  1

• latest model runner seems to have a few issues

  docker/model-runner#969 · Filter by author Glog78 opened on Jun 12, 2026
  4

• docker model package doesn't inherent model features ( no vision decoding )

  docker/model-runner#963 · Filter by author Glog78 opened on Jun 9, 2026
  1