List view

[CI]Enhance LLMC integration test

enhancementNew feature or requestNew feature or request

high priority

Status: Open.

Task

#1149 In intel/auto-round;

· yiliu30 opened on Dec 17, 2025

[Feature]: Add sageattentionV1 support

enhancementNew feature or requestNew feature or request

high priority

Status: Open.

#1503 In intel/auto-round;

· luoyu-intel opened on Mar 6, 2026

[Feature]: Support MXFP8/FP8 Model for ARK

enhancementNew feature or requestNew feature or request

Status: Open.

#1444 In intel/auto-round;

· luoyu-intel opened on Feb 12, 2026

xpu does not support fp8 model as input due to triton

Intel GPU

Status: Open.

Task

#827 In intel/auto-round;

· wenhuach21 opened on Sep 17, 2025

AutoScheme not supported in MLLM/VLM mode (Qwen3-VL) — timeline for support?

Status: Open.

#1273 In intel/auto-round;

· tosharybkin opened on Jan 13, 2026

low_cpu_mem_usage supports mxfp4/nvfp4

high priority

Status: Open.

#1127 In intel/auto-round;

· wenhuach21 opened on Dec 11, 2025

[Feature] Enhance hadamard transform

high priority

Status: Open.

#1569 In intel/auto-round;

· chensuyue opened on Mar 19, 2026

SGlang cannot load AutoRound model

bugSomething isn't workingSomething isn't working

Status: Open.

#982 In intel/auto-round;

· haofanwang opened on Nov 3, 2025

explore how a vLLM model can be used as input for quantization.

high priority

Status: Open.

Task

#1119 In intel/auto-round;

· wenhuach21 opened on Dec 10, 2025

decouple calibration code

Status: Open.

#810 In intel/auto-round;

· wenhuach21 opened on Sep 10, 2025

RuntimeError: Error(s) in loading state_dict for FP8Linear

Status: Open.

Bug

#1268 In intel/auto-round;

· chensuyue opened on Jan 13, 2026

[Feature]: Support Qwen3-TTS and Mimo-Audio quantization

enhancementNew feature or requestNew feature or request

Status: Open.

#1543 In intel/auto-round;

· WeiweiZhang1 opened on Mar 13, 2026

explore svdquant and EoRA

Status: Open.

Task

#947 In intel/auto-round;

· wenhuach21 opened on Oct 27, 2025

Add support for mamba models

Status: Open.

#374 In intel/auto-round;

· wenhuach21 opened on Dec 6, 2024

[Enhancement]: Refine AWQ backend selection: use compute_dtype as a filter criterion

enhancementNew feature or requestNew feature or request

Status: Open.

#1565 In intel/auto-round;

· WeiweiZhang1 opened on Mar 18, 2026

refine cuda kernel to huggingface/kernels

Status: Open.

Task

#837 In intel/auto-round;

· wenhuach21 opened on Sep 22, 2025

upstream AutoRound nvfp4 in LLMC

Status: Open.

Task

#1172 In intel/auto-round;

· xin3he opened on Dec 19, 2025

[Feature]: Implement model_free quantization method (W4A16）

enhancementNew feature or requestNew feature or request

high priority

Status: Open.

#1491 In intel/auto-round;

· xin3he opened on Mar 4, 2026

[Feature]: Packing format llm_compressor support INT4 (W4A16)

enhancementNew feature or requestNew feature or request

Status: Open.

#1567 In intel/auto-round;

· noobHappylife opened on Mar 19, 2026

[LLMC] Add DDP Support

enhancementNew feature or requestNew feature or request

high priority

llmc

Status: Open.

Task

#1408 In intel/auto-round;

· yiliu30 opened on Feb 5, 2026

[Feature]: Support AWQ based smoothing algorithm in W8A8, W4A4 scheme

enhancementNew feature or requestNew feature or request

high priority

Status: Open.

#1469 In intel/auto-round;

· thuang6 opened on Feb 25, 2026

customized dataset exception on LLMC side

bugSomething isn't workingSomething isn't working

high priority

llmc

Status: Open.

Task

#1080 In intel/auto-round;

· wenhuach21 opened on Nov 30, 2025

quant_lm_head=True actually didn't quant the lm head linear activation.

bugSomething isn't workingSomething isn't working

Status: Open.

#1044 In intel/auto-round;

· a32543254 opened on Nov 18, 2025

[Feature]: Support XPU INT2 kernels

enhancementNew feature or requestNew feature or request

Status: Open.

#1443 In intel/auto-round;

· luoyu-intel opened on Feb 12, 2026

[Upstream]: vLLM inference with AR kernels

high priority

Status: Open.

#1446 In intel/auto-round;

· luoyu-intel opened on Feb 12, 2026