Search results

[RFC]: Enable reproducible benchmarking in benchmark_serving_multi_turn with API-usage token counts

RFC

Status: Open.

#33640 In vllm-project/vllm;

· hyunnnchoi opened on Feb 3, 2026

[Bug]: DeepSeekV3.1 with fp8 kvcache in v0.15.0 produces garbled output

bugSomething isn't workingSomething isn't working

Status: Open.

#33638 In vllm-project/vllm;

· lyg95 opened on Feb 3, 2026

[Doc]: When running benchmarks, the prefix cache hit rate is abnormally high (not consistent with the production environment).

documentationImprovements or additions to documentationImprovements or additions to documentation

Status: Open.

#33632 In vllm-project/vllm;

· uOnePiece opened on Feb 3, 2026

[CI Failure]: Test Models (Qwen OMni)

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

Status: Open.

#33630 In vllm-project/vllm;

· robertgshaw2-redhat opened on Feb 3, 2026

[Bug]: Failed to run distributed inference with Gloo backend on aarch64

bugSomething isn't workingSomething isn't working

Status: Open.

#33628 In vllm-project/vllm;

· wenlujon opened on Feb 3, 2026

[Bug]: DeepSeek R1 with CUTLASS MLA Broken on B200

bugSomething isn't workingSomething isn't working

Status: Open.

#33627 In vllm-project/vllm;

· robertgshaw2-redhat opened on Feb 3, 2026

[Feature]: improve sleep mode to not break torch.cuda memory counters

feature requestNew feature or requestNew feature or request

Status: Open.

#33625 In vllm-project/vllm;

· stas00 opened on Feb 3, 2026

[Feature]: investigate if we can turn off the ahead-of-time autotuning block in vLLM

feature requestNew feature or requestNew feature or request

torch.compile

Status: Open.

#33622 In vllm-project/vllm;

· zou3519 opened on Feb 3, 2026

[Bug]: reasoning_content to reasoning rename caused regression in chat templates expecting reasoning_content

bugSomething isn't workingSomething isn't working

Status: Open.

#33616 In vllm-project/vllm;

· koush opened on Feb 2, 2026

[Bug] [CI, Build, and Release] Lost vllm against PyTorch nightly signal since Jan 28, 2026

bugSomething isn't workingSomething isn't working

Status: Open.

#33603 In vllm-project/vllm;

· atalman opened on Feb 2, 2026

vLLM replicas init failure when launching with Ray on a single instance

bugSomething isn't workingSomething isn't working

Status: Open.

#33601 In vllm-project/vllm;

· jiangwu300 opened on Feb 2, 2026

[RFC]: Improving vLLM Dependency Compatibility with Downstream Ecosystems

RFC

Status: Open.

#33599 In vllm-project/vllm;

· jeffreywang-anyscale opened on Feb 2, 2026