Skip to content

Pull requests: vllm-project/vllm

New pull request New
Clear current search query, filters, and sorts
5,371 Open 31,632 Closed
5,371 Open 31,632 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

[New Model][Nvidia] Add SM12x support for DeepSeek V4 Flash with essential fixes ci/build deepseek Related to DeepSeek models frontend kv-connector mrv2 Model Runner V2 specific needs-rebase new-model Requests to new models nvidia quantization speculative-decoding structured-output tool-calling v1
#41834 opened May 6, 2026 by jasl Contributor Loading…
465 🎉 4
[Core] Extensible (growable) KV cache ci/build cohere Related to Cohere models cpu Related to CPU backends deepseek Related to DeepSeek models documentation Improvements or additions to documentation DSv4 frontend glm gpt-oss Related to GPT-OSS models inkling intel-gpu Related to Intel GPU k3 kimi kv-cache-manager kv-connector llama Related to Llama models minimax mistral Related to Mistral models mrv2 Model Runner V2 specific multi-modality Related to multi-modality (#4194) needs-rebase nvidia performance Performance-related issues quantization qwen Related to Qwen models ray anything related with ray ready ONLY add when PR is ready to merge/full CI is needed rocm Related to AMD ROCm speculative-decoding structured-output tool-calling torch.compile
#50779 opened Aug 2, 2026 by njhill Member • Draft
38 🎉 2
Fix Mamba copy metadata high-bit pointers needs-rebase v1
#41995 opened May 7, 2026 by dipenbhuva Loading…
1
7 🎉 2
[Spec Decode][Hybrid] Add ngram-eagle SD method documentation Improvements or additions to documentation needs-rebase performance Performance-related issues speculative-decoding v1
#24344 opened Sep 5, 2025 by ekagra-ranjan Contributor Loading…
3 tasks done
@benchislett @WoosukKwon
50 🎉 2
Add DeepSeek-V4 DCP decode support deepseek Related to DeepSeek models DSv4 kv-cache-manager needs-rebase nvidia v1
#44573 opened Jun 4, 2026 by baonudesifeizhai Contributor Loading…
4 tasks
31 🎉 2
[MoE][Offload] Run MoE models exceeding VRAM via expert CPU offloading with GPU cache (--moe-expert-cache-size) ci/build documentation Improvements or additions to documentation frontend performance Performance-related issues quantization verified Run pre-commit for new contributors without triggering other tests
#37190 opened Mar 16, 2026 by e1n00r Loading…
70 🎉 1
[Perf][Kernel][ROCm] Add Triton split-KV paged decode fallback for gfx12 mrv1-only Issues/PRs which apply only to Model Runner V1 (not applicable to Model Runner V2) rocm Related to AMD ROCm v1
#45916 opened Jun 17, 2026 by feiyehua Loading…
4 tasks done
21 🎉 1
[ROCm] Fix two RDNA4 correctness bugs in the LLMM1 skinny-GEMM kernel performance Performance-related issues ready ONLY add when PR is ready to merge/full CI is needed rocm Related to AMD ROCm
#40827 opened Apr 24, 2026 by wjabbour Contributor Loading…
9 🎉 1
fix(kv-cache): allow TurboQuant on hybrid models quantization v1
#41123 opened Apr 28, 2026 by cderinbogaz Loading…
8 of 9 tasks
24 🎉 1
[EC Connector] Mooncake EC Connector for Distributed Encoder-Cache Transfer documentation Improvements or additions to documentation frontend kv-connector needs-rebase v1
#40695 opened Apr 23, 2026 by fake0fan Contributor Loading…
11 🎉 1
[ROCm][Bugfix] Densify strided activations for wvSplitKQ bug Something isn't working rocm Related to AMD ROCm
#50618 opened Jul 31, 2026 by JohnQinAMD Contributor Loading…
5 🎉 1
feat(openai): add per-request timing metrics and completion_tokens_de… documentation Improvements or additions to documentation frontend needs-rebase
#36383 opened Mar 8, 2026 by dryu-nv • Draft
19 🎉 1
[WIP][Proof of concept] Overlap model loading and torch.compile frontend needs-rebase v1
#36072 opened Mar 5, 2026 by zou3519 Collaborator • Draft
9 🎉 1
[Bugfix] Handle None compilation times in v1 Executor initialize_from_config bug Something isn't working needs-rebase stale Over 90 days of inactivity v1
#37628 opened Mar 20, 2026 by ec-jt Loading…
2 🎉 1
[Feature] Support batch invariance on ROCm ci/build needs-rebase nvidia quantization rocm Related to AMD ROCm
#52231 opened Aug 14, 2026 by mawong-amd Contributor Loading…
4 tasks
21 🎉 1
feat: FlashInfer CuteDSL MegaMoE integration deepseek Related to DeepSeek models DSv4 mrv2 Model Runner V2 specific nvidia quantization
#54049 opened Aug 27, 2026 by jdebache Contributor Loading…
7 🎉 1
[Bugfix][AsyncScheduler] Reserve placeholders for the final prefill chunk of a streaming update bug Something isn't working needs-rebase scheduler streaming-input
#50692 opened Aug 1, 2026 by Aayush7352 Loading…
1
2
[ROCm][AITER] Add support for new per-stream workspace requirement by AITER v0.1.20 mrv2 Model Runner V2 specific needs-rebase nvidia rocm Related to AMD ROCm
#54095 opened Aug 27, 2026 by qli88 Contributor • Draft
1
[ROCm] optimize memory for GLM 5.2 glm needs-rebase rocm Related to AMD ROCm v1
#50459 opened Jul 30, 2026 by Concurrensee Contributor • Draft
1
[Rust Frontend] Harden resumable session lifecycle in engine-core-client rust streaming-input
#53454 opened Aug 23, 2026 by TanNgocDo Contributor Loading…
2 tasks done
[Perf][PCP] Shard decode requests across PCP ranks deepseek Related to DeepSeek models mrv2 Model Runner V2 specific needs-rebase ready ONLY add when PR is ready to merge/full CI is needed
#52162 opened Aug 13, 2026 by pisceskkk Contributor Loading…
4 tasks done
6
[Bugfix][Spec Decode] DFlash2: accept unquantized linear LM heads in the candidate selector bug Something isn't working dflash mrv2 Model Runner V2 specific needs-rebase new-model Requests to new models quantization qwen Related to Qwen models speculative-decoding
#52883 opened Aug 19, 2026 by oceanplexian Loading…
4
[ROCm][aiter] officially supporting aiter Triton kernels+dsv4 on RDNA3 DSv4 quantization rocm Related to AMD ROCm torch.compile
#52970 opened Aug 19, 2026 by amd-xavierwang Loading…
2
[Bugfix] Format kernel-import errors eagerly so warning_once does not retain them bug Something isn't working cpu Related to CPU backends nvidia
#54098 opened Aug 27, 2026 by zwang86 Loading…
1
4
[Perf] Add tuning configs for SM103 Kimi k3 nvfp4 DEP projection GEMM k3 kimi needs-rebase
#53794 opened Aug 25, 2026 by wzhao18 Contributor • Draft
4 tasks
7
ProTip! Add no:assignee to see everything that’s not assigned.