Issues

• [Feature] Support pending external linker lookup at native prefill admission

  sgl-project/sglang#44872 · Filter by author feichai0017 opened on Oct 17, 2026

• [Feature] Prepare registered radix cache backends before worker graph initialization

  sgl-project/sglang#44855 · Filter by author feichai0017 opened on Oct 17, 2026

• [Bug] pause_generation(abort) + release_memory_occupation right after a client disconnect crashes scheduler

  sgl-project/sglang#44840 · Filter by author Yunzez opened on Oct 17, 2026

• [Bug] With Speculative decoding + abort, completion can return more tokens than max_tokens

  sgl-project/sglang#44839 · Filter by author Yunzez opened on Oct 17, 2026

• [Bug] With pause_generation(in_place) then pause_generation(abort), the server hangs until killed by watchdog

  sgl-project/sglang#44835 · Filter by author Yunzez opened on Oct 17, 2026

• [Bug] update_weights_from_disk after release_memory_occupation is accepted and crashes the scheduler (segfault in weight_loader)

  sgl-project/sglang#44834 · Filter by author Yunzez opened on Oct 17, 2026

• [CI] Post-merge runs on main cancel each other in NPU/XPU/Xeon/MUSA/SMG workflows

  sgl-project/sglang#44832 · Filter by author JerryChaox opened on Oct 17, 2026

• [Bug] DeepSeek-V4 BF16 checkpoint is routed to the MXFP4 expert path and fails in fused MoE

  sgl-project/sglang#44810 · Filter by author micahtyong opened on Oct 16, 2026
  2

• [Bug] [AMD] select_experts casts the fp32 MoE correction bias to bf16 before #37133's fp32 aiter path, so GLM-5.x still routes with a bf16 bias

  sgl-project/sglang#44794 · Filter by author jhinpan opened on Oct 16, 2026
  2

• [Bug] MiMo-V2.6: MiMoV2Processor hard-codes CLIP mean/std, checkpoint preprocessor_config specifies ImageNet

  sgl-project/sglang#44789 · Filter by author bjolo opened on Oct 16, 2026
  1

• [RFC / Feature] SPICY-KV: 28.6x–50x Pre-RoPE Algebraic Multigrid KV Compression for RadixAttention & Long-CoT (91.7%–100% RULER Recall & 100% Shared-Prefix Multi-Query Reuse)

  sgl-project/sglang#44786 · Filter by author Yuge-Y opened on Oct 16, 2026

• [Bug] Kimi-K3: nvidia_kda prefill produces NaN after ~1,800 tokens (Triton is correct)

  sgl-project/sglang#44758 · Filter by author Venkat2811 opened on Oct 16, 2026
  1

• Quick survey: your experience as a PyTorch Ecosystem maintainer

  sgl-project/sglang#44754 · Filter by author michellearoth opened on Oct 16, 2026

• [Bug] Grayscale JPEG's processing failes on normalization step

  sgl-project/sglang#44748 · Filter by author kalashnikova04 opened on Oct 16, 2026

• [Bug] When client disconnect with pause_generation(mode="abort") and release_memory_occupation kills the scheduler

  sgl-project/sglang#44745 · Filter by author Yunzez opened on Oct 16, 2026

• [Bug] pause_generation(mode="retract") while a request's grammar is compiling, then release_memory_occupation with an AssertionError kills the scheduler

  sgl-project/sglang#44744 · Filter by author Yunzez opened on Oct 16, 2026

• [RFC] scheduler: Overlap prefill scheduling and KV loading with model execution

  sgl-project/sglang#44721 · Filter by author ttaohe opened on Oct 16, 2026

• [RFC] MoE routing simulator for prefill imbalance profiling: balanced reference and GPU-faithful replay

  sgl-project/sglang#44697 · Filter by author Phoenix3334 opened on Oct 16, 2026

• [Roadmap] Muse Spark Open Development & Optimization

  sgl-project/sglang#44678 · Filter by author Fridge003 opened on Oct 16, 2026

• [Roadmap] Intel CPU (Xeon) Backend Optimization (2026Q4)

  cpuintelroadmap
  sgl-project/sglang#44675 · Filter by author mingfeima opened on Oct 16, 2026

• [Roadmap] Intel GPU (XPU) Backend Optimization (2026Q4)

  intelroadmapxpu
  sgl-project/sglang#44674 · Filter by author mingfeima opened on Oct 16, 2026

• [Perf] HiCache Throughput Loss Analysis in Long-Context MLA Workloads

  sgl-project/sglang#44656 · Filter by author githlangub opened on Oct 16, 2026
  2

• [Bug] Client-declared mm_hashes controls multimodal prefix-cache identity: wrong-image KV reuse and cross-tenant image disclosure

  sgl-project/sglang#44641 · Filter by author yyymk opened on Oct 16, 2026
  2

• [RFC] Context-parallel support for the unified L3 key scheme

  sgl-project/sglang#44635 · Filter by author momaek opened on Oct 16, 2026
  3

• [Bug] Out-of-vocab stop_token_ids crash the server in the min_new_tokens penalizer

  sgl-project/sglang#44535 · Filter by author KrishnanPrash opened on Oct 15, 2026