Search results

[Feature]: Support responses api in tokenize endpoint

feature requestNew feature or requestNew feature or request

Status: Open.

#40556 In vllm-project/vllm;

· andrewbcohere opened on Apr 21, 2026

[AMD][CI Failure][Tracker] Static dashboard tracker for current CI failures

ci-failureIssue about an unexpected test failure in CIIssue about an unexpected test failure in CI

rocmRelated to AMD ROCmRelated to AMD ROCm

Status: Open.

#40554 In vllm-project/vllm;

· AndreasKaratzas opened on Apr 21, 2026

[Bug]: Worse EAGLE3 acceptance rates on MRV2

bugSomething isn't workingSomething isn't working

help wantedExtra attention is neededExtra attention is needed

Status: Open.

#40551 In vllm-project/vllm;

· benchislett opened on Apr 21, 2026

[Feature]: Integrate fused kMoEFinalizeARResidualRMSNorm from FlashInfer

feature requestNew feature or requestNew feature or request

help wantedExtra attention is neededExtra attention is needed

Status: Open.

#40544 In vllm-project/vllm;

· benchislett opened on Apr 21, 2026

[Feature]: Dynamic PDL Enablement

feature requestNew feature or requestNew feature or request

Status: Open.

#40543 In vllm-project/vllm;

· benchislett opened on Apr 21, 2026

[Feature]: Support iterative in-place weight editing on TP workers (online RLHF / steering / abliteration)

Status: Open.

#40536 In vllm-project/vllm;

· wuwangzhang1216 opened on Apr 21, 2026

[Bug]: Weight loading for meta-llama/Llama-4-Scout-17B-16E-Instruct is extremely slow

bugSomething isn't workingSomething isn't working

Status: Open.

#40535 In vllm-project/vllm;

· bnellnm opened on Apr 21, 2026

[RFC]: Hybrid checkpoint ABI for non-KV prefix resume

RFC

Status: Open.

#40533 In vllm-project/vllm;

· CarlOskarRost opened on Apr 21, 2026

[Bug]: Forced tool_choice crashes with AssertionError when reasoning_parser consumes all content

Status: Open.

#40528 In vllm-project/vllm;

· liuchenbing2026 opened on Apr 21, 2026

[Bug]: Streaming output incorrectly mapped to reasoning field instead of content when enable_thinking=False

bugSomething isn't workingSomething isn't working

Status: Open.

#40466 In vllm-project/vllm;

· linqiuu opened on Apr 21, 2026

[RFC]: Convert Triton kernels from raw pointers to block pointers

RFC

Status: Open.

#40458 In vllm-project/vllm;

· ohadeytan opened on Apr 21, 2026

[Bug]: Per-attention-head quantization is currently available only with the Flash Attention backend and requires the calibration pathway provided by llm-compressor.

bugSomething isn't workingSomething isn't working

Status: Open.

#40444 In vllm-project/vllm;

· cuihangbin opened on Apr 21, 2026