Issues

• [Re-land] #118 jemm native worker: 9 deduped commits onto main, conflict map and fidelity check

  InferMatrixLab/system1-omni#154 · Filter by author hsliuustc0106 opened on Oct 21, 2026
  1

• [Triage] 30 open PRs: merge / rebase / author buckets, merge order and acceptance criteria (2026-10-21)

  InferMatrixLab/system1-omni#153 · Filter by author hsliuustc0106 opened on Oct 21, 2026
  5

• [RFC]: Next-phase roadmap: prefill-only execution, reuse and scheduling

  InferMatrixLab/system1-omni#150 · Filter by author twu3202 opened on Oct 19, 2026

• [New Model]: Support the Vela 2.0 series

  new model
  InferMatrixLab/system1-omni#137 · Filter by author MrDongsls opened on Oct 11, 2026
  

• [New Model]: Support the D3 / Decision 3.0 series — community help wanted

  help wantednew model
  InferMatrixLab/system1-omni#136 · Filter by author hsliuustc0106 opened on Oct 11, 2026
  1

• [New Model]: Support internlm/Intern-Decision (4B / 2B / 0.8B) — third Qwen3.5 decision family, single-prefill multi-field readout

  help wantednew model
  InferMatrixLab/system1-omni#133 · Filter by author hsliuustc0106 opened on Oct 10, 2026
  1

• [Benchmark] Preserve complete serving latency evidence and offline summaries

  InferMatrixLab/system1-omni#130 · Filter by author linear3735 opened on Oct 10, 2026

• [Feature]: Laya recipe: move laya[serve] from 0.3.20 to 0.3.28 so a request's max_len / head_max_len take effect

  enhancement
  InferMatrixLab/system1-omni#128 · Filter by author chaimaerachdi opened on Oct 9, 2026
  1

• [Tracking]: System1-Omni roadmap: models, features, and agent integration

  InferMatrixLab/system1-omni#125 · Filter by author hsliuustc0106 opened on Oct 8, 2026

• [RFC]: Native Rust/CUDA support for MaestroYan/JEMM and community help

  help wantednew modelRFC
  InferMatrixLab/system1-omni#121 · Filter by author hsliuustc0106 opened on Oct 8, 2026
  1

• [RFC]: Native Rust/CUDA support for OmniJev-4B v1.1 and community help

  help wantednew modelRFC
  InferMatrixLab/system1-omni#114 · Filter by author hsliuustc0106 opened on Oct 7, 2026
  5

• [RFC]: Support openjev/openjev in System1-Omni — community help wanted

  help wantednew modelRFC
  InferMatrixLab/system1-omni#95 · Filter by author hsliuustc0106 opened on Oct 6, 2026
  1

• [Feature]: Reproducible Open-Jev agent demo — community help wanted

  enhancementhelp wanted
  InferMatrixLab/system1-omni#87 · Filter by author hsliuustc0106 opened on Oct 5, 2026
  1

• [RFC]: Reuse shared observations across decision questions — community help wanted

  help wantedperformanceRFC
  InferMatrixLab/system1-omni#85 · Filter by author hsliuustc0106 opened on Oct 5, 2026
  6

  

• [New Model]: Add Valen-Preview-0923 reference serving — community help wanted

  help wantednew model
  InferMatrixLab/system1-omni#84 · Filter by author hsliuustc0106 opened on Oct 5, 2026
  6

  

• [Tracking]: Model support status

  documentation
  InferMatrixLab/system1-omni#83 · Filter by author hsliuustc0106 opened on Oct 5, 2026

• [Performance]: Optimize native Rust/CUDA Laya inference

  help wantedperformance
  InferMatrixLab/system1-omni#66 · Filter by author hsliuustc0106 opened on Oct 3, 2026
  1

• [RFC]: Write down which parts of the /v1/systemone contract all workers share

  InferMatrixLab/system1-omni#61 · Filter by author twu3202 opened on Oct 1, 2026
  2

• [RFC]: Native Rust/CUDA inference for Mapika/decider-2b

  RFC
  InferMatrixLab/system1-omni#57 · Filter by author hsliuustc0106 opened on Oct 1, 2026

• [Performance]: Reproducible Laya benchmarks against the original runtime

  help wantedperformance
  InferMatrixLab/system1-omni#39 · Filter by author hsliuustc0106 opened on Sep 29, 2026

• Cua-S1 multimodal stack: validation results and cross-PR integration checklist

  InferMatrixLab/system1-omni#36 · Filter by author Levius-Fubuki opened on Sep 29, 2026
  2

• [New Model]: LFM2.5-350M choice execution

  InferMatrixLab/system1-omni#31 · Filter by author prettygirlisnotme opened on Sep 28, 2026

• [New Model]: Support Kev (0.5B–9B) — second Qwen3.5 decision model, reusing the qwen3_5 CUDA backend

  InferMatrixLab/system1-omni#27 · Filter by author xiaoyu-xyz opened on Sep 28, 2026

• [RFC]: Native Laya inference with Rust and CUDA

  InferMatrixLab/system1-omni#14 · Filter by author linear3735 opened on Sep 28, 2026
  1

• [New Model]: Support Cua-S1 4B 0.2 — community help wanted

  help wantednew model
  InferMatrixLab/system1-omni#10 · Filter by author hsliuustc0106 opened on Sep 27, 2026
  8