Issues

• Qwen3VL export has inconsistent dtype for visual_encoder

  onnxruntime/mobius#750 · Filter by author vigneshm2 opened on Sep 30, 2026

• Add a CPU runtime contract test for mixed-width QMoE

  onnxruntime/mobius#745 · Filter by author titaiwangms opened on Sep 22, 2026
  

• Add Qwen3.8 packed hybrid PagedAttention export for ORT GenAI Engine

  aienhancementtracking
  onnxruntime/mobius#737 · Filter by author titaiwangms opened on Sep 16, 2026
  

• Support per-layer Olive mixed-precision QMoE export

  onnxruntime/mobius#735 · Filter by author titaiwangms opened on Sep 14, 2026
  

• Track official VibeVoice 1.5B HF sidecars

  tracking
  onnxruntime/mobius#727 · Filter by author justinchuby opened on Sep 4, 2026

• [Model Request]: Enable Qwen4Exp GGUF payload import and execution

  model-request
  onnxruntime/mobius#699 · Filter by author justinchuby opened on Aug 29, 2026

• Clarify GGUF quantized storage fidelity and BitShift fallback compute

  onnxruntime/mobius#667 · Filter by author justinchuby opened on Aug 27, 2026

• Close remaining heterogeneous recurrent-state runtime gaps

  onnxruntime/mobius#605 · Filter by author justinchuby opened on Aug 25, 2026

• [Model Request]: nvidia/magpie_tts_multilingual_357m

  model-request
  onnxruntime/mobius#570 · Filter by author justinchuby opened on Aug 23, 2026

• 4bit diffusion models

  onnxruntime/mobius#569 · Filter by author justinchuby opened on Aug 23, 2026

• Weekly prompt

  onnxruntime/mobius#481 · Filter by author justinchuby opened on Aug 14, 2026

• Convert weather next

  onnxruntime/mobius#468 · Filter by author justinchuby opened on Aug 10, 2026
  

• Plans to Support VibeVoice/Nemotron - TTS Support

  onnxruntime/mobius#356 · Filter by author leestott opened on Jun 15, 2026
  2

  

• Emit opset-24 external KV cache (nonpad_kv_seqlen) for bias decoders to recover near-GQA decode (Gemma-4, sliding-window)

  onnxruntime/mobius#349 · Filter by author justinchuby opened on Jun 9, 2026
  1

  

• Benchmark: Gemma4 ORT GenAI inference — f16 vs Q4_K_M vs NF4 (CUDA + CPU)

  onnxruntime/mobius#346 · Filter by author justinchuby opened on Jun 8, 2026
  1

• Emit maskless is_causal=1 + nonpad_kv_seqlen static-cache graph once upstream lands

  onnxruntime/mobius#345 · Filter by author titaiwangms opened on Jun 8, 2026
  

• Duplicated badges still being reported

  onnxruntime/mobius#344 · Filter by author tadani3 opened on Jun 8, 2026
  2

  

• Not tested model logic is still incorrectly reported in Dashboard

  onnxruntime/mobius#343 · Filter by author tadani3 opened on Jun 8, 2026
  

• Count Locally Verified Models as Passes in Mobius Dashboard

  onnxruntime/mobius#342 · Filter by author tadani3 opened on Jun 8, 2026
  

• Validate QDQ with gemma 4

  onnxruntime/mobius#332 · Filter by author justinchuby opened on Jun 4, 2026
  

• Public API design proposal: promote curated surface, demote internals to _internal

  onnxruntime/mobius#325 · Filter by author justinchuby opened on Jun 1, 2026

• Gemma 4 Benchmark Reproduction & Olive Eval Integration Plan

  onnxruntime/mobius#285 · Filter by author justinchuby opened on May 7, 2026
  13

• Gemma4 ONNX + CUDA support: all PRs tracker

  onnxruntime/mobius#257 · Filter by author justinchuby opened on May 5, 2026
  6

• Add Nemotron streaming - 0.6 B support for Mobius

  onnxruntime/mobius#184 · Filter by author rui-ren opened on Apr 21, 2026
  2

  

• Follow-ups from PR #130: golden reference JSON + LazyTensor FP8 dequant

  onnxruntime/mobius#156 · Filter by author titaiwangms opened on Apr 10, 2026
  5