Bump vllm from 0.7.3 to 0.26.0

[dependabot]

Bumps vllm from 0.7.3 to 0.26.0.

Release notes

Sourced from vllm's releases.

v0.26.0

vLLM v0.26.0 Release Notes

Highlights

This release features 411 commits from 212 contributors (61 new)!

• New Inkling model family with a full support stack: base modeling (#48799), piecewise CUDA graph support (#48822), Hopper FA4 relative attention (#48858), MTP=1 speculative decoding (#48869), LoRA (#48884), and standard ModelOpt NVFP4 quantization (#48990).
• DeepSeek-V4 performance push across vendors: a specialized routing kernel (2.94% E2E TPOT, #48660), fused_topk_bias (1.5–2x kernel, #47463), and redundant repeat/copy removal (1.8% E2E TPOT, #48137), plus ROCm two-stage compressor for HCA prefill (#47718), sparse decode/prefill optimizations (#48519, #48788, #46275), and DSpark speculative decoding on AMD (#47419) and XPU (#47677).
• fp32 lm_head for generation models via head_dtype (#48390), extended to the LoRA path (#48525) and given a ROCm torch.mm fast path (#48688), improving accuracy for generation heads.
• Flexible attention backends: the attention backend can now be selected per KV-cache group (#48012), and sliding-window support is now an explicit backend capability (#48011) — improving support for hybrid models.
• KV offloading & tiered secondary storage matured substantially: offloading metrics (#45958, #47666, #47679), tier-owned event handling (#46544, #47923), object-store secondary tier with workload identity (#47063, #47274, #48150), DP-replica-aware tiering (#47987), and encoder-cache (EC) connectors including CPU offloading (#42433, #47423).
• Rust frontend gained multimodal video (#47959) and audio (#48554), a Seed-OSS tool parser (#47741), and a native vllm-bench port (#48107).
• Transformers 5.13.0 (#47867) with more models migrated to the Transformers modeling backend: Olmo/Olmo2 (#48100), MistralLarge3 (#48153), and HunyuanVL (#47872).

Model Support

• New models: Inkling family (#48799, #48822, #48858, #48869, #48884, #48990), BertForMaskedLM (#48463), RobertaForTokenClassification / XLMRobertaForTokenClassification (#47991), LongCat-Flash-Lite n-gram embedding (#47857), Cosmos3 Edge Reasoner (#48291) and Cosmos3-Super registration (#48211), TranslateGemma-12b-it (#41599).
• Transformers backend migrations: Olmo/Olmo2 (#48100), MistralLarge3 to AutoWeightsLoader (#48153), HunyuanVL native transformers processor for transformers 5.13 (#47872).
• GLM5.2: migrate MoE sequence-parallel support to the non-torch-compiled path (#47881).
• LoRA: FlashInfer MoE LoRA for BF16 models (#48632), LoRA for tower/connector in LlavaNextVideo (#48594), fp32 lm_head on the LoRA path (#48525), optimized TrtLlmLoRAExperts (#48759).
• Multimodal: automatic fallback to ViT data parallelism when TP is unavailable (#49046).
• Fixes: correct pooling scores for chunked prefill under torch.compile (#48901).

Engine Core

• fp32 lm_head for generation models via head_dtype (#48390); lower memory for capturing large CUDA graph sizes (#48483); opt-in persistence and reuse of the memory-profiling result across boots (#47388); improved InstantTensor loading (#46868).
• Attention: select a different attention backend per KV-cache group (#48012); sliding-window as an explicit backend capability (#48011); KV-cache layout refactor packing K/V into the content dim across backends (#44455); MRV2 virtual-batch PCP for MLA (#46570).
• Speculative decoding: runtime draft weight update (#46725), hybrid (SWA + full attention) DFlash drafters (#47914), SWA support for qwen-eagle3 (#47568), Gemma4-12B DSpark draft model (#47216), DSv4 DSpark on AMD (#47419), separate kv_cache_dtype for speculative_config (#48787).
• KV offloading: basic offloading metrics (#45958), split CPU cache usage into read/write gauges (#47666) and tiering-lookup-delay into sync/async histograms (#47679), tier-owned event handling and BlockStored events (#46544, #47923), object-store secondary tier with workload identity (#47063, #47274, #48150), DP-replica-aware tiering (#47987), blocks_per_chunk config for heterogeneous KV groups (#48878), P2P default host/port env vars (#47636).
• Caching: partial prefix-cache hit for hybrid models (#46384), selective hybrid cache retention (#47782), report prefix-cache-reused blocks in full report mode (#45261).
• Reasoning: optimize TPOT for thinking budget when used with speculative decoding (#46662).
• RLHF: stateful trainer-send abstractions (#48042).
• Fixes: host memory leak from undrained new_block_ids (#44490), DSv3.2 + MTP + sequence-parallel accuracy (#48036).

Hardware & Performance

• DeepSeek-V4: specialized routing kernel (2.94% E2E TPOT, #48660), fused_topk_bias 1.5–2x (#47463), redundant repeat/copy removal (1.8% TPOT, #48137).
• MoE router GEMMs: BF16x3 router GEMM (#47973), FP32 router GEMV (#48335), generic CuteDSL LL BF16 router GEMM (#42562); TRTLLM BF16 MoE modular kernel (#45182); write FlashInfer combine into final output (#47156).
• Qwen: fuse more RMSNorm + all-reduce in Qwen3.5 (#46998), replace MoE all-reduce with reduce-scatter (#47006), Qwen3.5 H20 optimization (#48350), expand Triton warmup coverage (#47546).
• MLA: dense MHA path for short sparse-MLA sequences (#47327); MiniMax-M3 long-context decode indexer on sm100 (#48582).
• Kernels: CUDA kernel for ReLUSquaredActivation / relu^2 (#39058), Helion kernel lazy registration (#48264), vectorize _copy_mamba_state_block to uint64 (#48110), stop upcasting logits to fp32 in the sampler (#48641).
• ROCm: fp32 head_dtype torch.mm fast path (#48688), DSv4 two-stage compressor kernel (#47718), sparse decode/prefill optimizations (#48519, #48788, #46275), DSv3.2 sparse MLA KV-split heuristic (#46832) and MTP CUDA-graph mode (#45149), MXFP8 GEMM for MiniMax-M3 (#46117), AITER sparse paged attention + spec decode for MiniMax-M3 (#47287, #47984), MiniMax-M2 fused QK-norm + all-reduce via AITER (#44849), HybridW4A16 linear kernel (#40977), Qwen3-30B-A3B QK-Norm+RoPE+KV runtime fusion (#42749).
• XPU: batch-invariant kernels (#41934), HND KV layout support (#47975), DSpark spec decode for DSv4 (#47677), nightly/release image publishing (#47880, #48126).
• CPU: DFlash speculative decoding for GDN models on CPU (#46090), s390x NUMA topology (#40714), native macOS arm64 CPU wheel builds (#48289); POWER VSX math function optimization (#47321) and IBM Power docker builds using prebuilt wheels (#46017).
• Distributed fusion: FlashInfer MNNVL all-reduce RMS quant fusion (#48064).
• Build/autotune: arm64 Blackwell SM10x/SM110 image builds (#48041); skip CuTeDSL fp4_gemm autotuning by default (#48268).

Large Scale Serving & Distributed

• Decode Context Parallel (DCP): hybrid attention support (#40996), DCP + Eagle for Tokenspeed MLA backends (#48180).
• PD disaggregation: NIXL pipeline-parallel prefill in push mode (#45880).
• Encoder-cache connectors: EC transfer params (#42433) and CPU-offloading EC connector (#47423).

... (truncated)

Commits

• 568afb3 [CI/Build] Refresh tags before building macOS wheel (#49901)
• f265493 [ROCm] [Release] [Bugfix] Fix the per commit wheel release pipeline. (#49245)
• ffd46bf [Bugfix] Register axk1 config to fix A.X-K1 init (#49727)
• ffd6ee4 [CI] Bump PyTorch Compilation Unit Tests timeout to 150 min (#49606)
• bb26ce8 [CI] Increase timeout of pytorch-compilation-unit-tests (#49450)
• 091db8b [CI] Increase timeouts for jobs exceeding current limits (#49374)
• ba694b8 [CI] Bump timeout of entrypoints-integration-api-server-openai-part-2 (#49359)
• e5949f1 [Bugfix] handle grammar compilation failures to avoid engine crash (#47312)
• 8b30569 [Bugfix] Fix DeepGEMM warmup when using `FlashInferFp8DeepGEMMDynamicBlockSca...
• 9d37a50 [Bugfix][Attention] Ignore empty MLA context chunks during merge (#49294)
• Additional commits viewable in compare view

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

---

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

• @dependabot rebase will rebase this PR
• @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
• @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
• @dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)
  You can disable automated security fix PRs for this repo from the Security Alerts page.

[entelligence-ai-pr-reviews]

EntelligenceAI PR Summary

Updates the pinned vLLM dependency in pyproject.toml from 0.7.3 to 0.26.0, incorporating the upstream release’s new features and fixes. No application code or configuration beyond the dependency version was changed.

---

Review Scorecard

Dimension	Rating	Basis
Code Quality	●●●●● 5/5 — Excellent	no critical/significant findings — rated Excellent
Blast Radius	Low	changed symbols are referenced only within their own file(s); no high-impact surface touched, 1 file(s) / ~2 line(s) changed (size only — not a blast signal)
Merge Confidence	●●●●● 5/5 — Safe to Merge	code quality 5/5 × Low blast radius

Safe to merge — this is a clean, low-blast-radius dependency-only change. pyproject.toml updates the pinned vLLM version without altering application code, configuration, or runtime logic, and no correctness, security, robustness, or convention issues were identified. There are no pre-existing unresolved issues relevant to this PR.

Key Findings:

• The change is isolated to the vLLM version pin in pyproject.toml, with no unrelated edits.
• No application code paths, error handling, access control, concurrency behavior, or data integrity logic were modified.

Evaluated against

• 1/1 changed files reviewed
• criteria: correctness, security & access control, robustness & error handling, concurrency & data integrity, repo conventions / steering docs
• steering docs: none found in repo

Files requiring special attention

• pyproject.toml

[entelligence-ai-pr-reviews]

Walkthrough

Updates the project’s pinned vLLM dependency from 0.7.3 to 0.26.0, upgrading the vLLM runtime used by the application and potentially requiring compatibility updates for integrated APIs or behavior.

Changes

File(s)	Summary
pyproject.toml	Updated the pinned vLLM dependency from version 0.7.3 to 0.26.0.

Sequence Diagram

This diagram shows the interactions between components:

sequenceDiagram
    title vLLM Dependency Upgrade and Inference Flow

    participant Build as "Build / Package Manager"
    participant Config as "pyproject.toml"
    participant App as "Application"
    participant VLLM as "vLLM Runtime 0.26.0"
    participant Model as "Loaded Language Model"

    Build->>Config: Read dependency specification
    Config-->>Build: Resolve vLLM==0.26.0
    Build->>VLLM: Install and make runtime available

    App->>VLLM: Initialize inference engine
    VLLM->>Model: Load model
    Model-->>VLLM: Model ready
    VLLM-->>App: Inference engine initialized

    loop For each inference request
        App->>VLLM: Submit prompt and generation parameters
        VLLM->>Model: Generate tokens
        Model-->>VLLM: Generated output
        VLLM-->>App: Return inference result
    end

    Note over Config, VLLM: The PR changes the pinned vLLM version from 0.7.3 to 0.26.0.

Loading

[dependabot]

Superseded by #11.