● PyPI package
vllm
Vllm has 113 security advisories. The newest release any of them names as the fix is 0.32.0.
As of , vllm has 113 security advisories. The newest release any of them names as the fix is 0.32.0.
Your version, and the version to upgrade to
| If you run | Upgrade to | Advisories |
|---|---|---|
| Before 0.32.0 | 0.32.0 | 113 advisories |
Security advisories, newest first
| Advisory | What it is | Affects | Fixed in | Published |
|---|---|---|---|---|
| CVE-2026-115913GHSA-8397-8qf6-f2gm | vLLM: Phi-4 multimodal audio encoder materializes and retains full attention matrices: a single max-length audio OOM-crashes the vLLM engineModerate severity | Before 0.32.0 | 0.32.0 | |
| CVE-2026-115914GHSA-gq5q-89c7-vhvx | vLLM DeepSeek-OCR image processing vulnerable to denial of serviceModerate severity | From 0.11.1 before 0.32.0 | 0.32.0 | |
| CVE-2026-114053GHSA-h3rc-6mm3-gc2m | vLLM: Request-controlled mm_processor_kwargs.code_revision allows remote code executionHigh severity | From 0.7.3 before 0.31.0 | 0.31.0 | |
| CVE-2026-115918GHSA-x567-49g5-p4cf, PYSEC-2026-4349 | vLLM: Responses API bypasses trust_request_chat_template for chat_template_kwargsModerate severity | From 0.16.0 before 0.32.0 | 0.32.0 | |
| CVE-2026-115917GHSA-vvpg-rrww-wpmh, PYSEC-2026-4348 | vLLM: Permissive default CORS origins allow cross-origin requests to the vLLM HTTP serverLow severity | Before 0.32.0 | 0.32.0 | |
| CVE-2026-115916GHSA-q97c-p368-75c4, PYSEC-2026-4347 | vLLM: Out-of-vocabulary prompt token id passes input validation and triggers a CUDA device-side assert (engine crash / DoS)Moderate severity | Before 0.32.0 | 0.32.0 | |
| CVE-2026-115915GHSA-h3rc-j82p-2c3h, PYSEC-2026-4346 | vLLM: Untrusted EPS inputs can trigger denial of service or Ghostscript RCEHigh severity | Before 0.32.0 | 0.32.0 | |
| CVE-2026-115912GHSA-6757-hggm-hp2w, PYSEC-2026-4345 | vLLM: Public multimodal SHM handle forgery reaches worker-side `pickle.loads()` with insecure serialization disabledHigh severity | Before 0.32.0 | 0.32.0 | |
| CVE-2026-115911GHSA-542m-9234-8gvf, PYSEC-2026-4344 | vLLM: DoS via unbounded chat_template Jinja2 rendering exhausts the renderer thread poolModerate severity | From 0.19.0 before 0.32.0 | 0.32.0 | |
| CVE-2026-116178PYSEC-2026-4353 | vLLM from 0.20.0 before 0.30.0 contains a denial of service vulnerability that allows remote attackers to crash the engine by sending inconsistent multimodal metadata to the disaggregated /inference/v1/generate endpoint. | From 0.20.0 before 0.30.0 | 0.30.0 | |
| CVE-2026-116177PYSEC-2026-4352 | vLLM from 0.10.2 before 0.25.0 contains an uncontrolled resource consumption vulnerability that allows unauthenticated attackers to exhaust memory because the Nemotron-VL processors globally set Image.MAXIMAGEPIXELS to None. | From 0.10.2 before 0.25.0 | 0.25.0 | |
| CVE-2026-116176PYSEC-2026-4351 | vLLM from 0.9.2 before 0.24.0 contains a server-side request forgery vulnerability that allows API callers to force outbound connections by embedding a decode address in the X-Request-Id header. | From 0.9.2 before 0.24.0 | 0.24.0 | |
| CVE-2026-116173PYSEC-2026-4350 | vLLM from 0.22.0 before 0.24.0 contains a denial of service vulnerability that allows remote attackers to terminate EngineCore by bypassing logprobtokenids limits in the Rust gRPC frontend. | From 0.22.0 before 0.24.0 | 0.24.0 | |
| CVE-2026-114055GHSA-gx7p-2j49-hfq4 | vLLM : DoS via unvalidated mm_processor_kwargs in multimodal image processors (Isaac, InternVL, H2OVL, etc.)High severity | From 0.7.3 before 0.31.0 | 0.31.0 | |
| CVE-2026-114050GHSA-p92p-rxj5-7p2x | vLLM: Forged multimodal UUID + P0/P1 processor-cache eviction drift: unauthenticated remote crash of the whole vLLM engineModerate severity | Before 0.31.0 | 0.31.0 | |
| CVE-2026-114054PYSEC-2026-4362 | vLLM before 0.31.0 contains a cache aliasing vulnerability that allows authenticated attackers to collide shared cache keys by supplying caller-controlled values such as EXIF ImageID or cachesalt. | Before 0.31.0 | 0.31.0 | |
| CVE-2026-114052PYSEC-2026-4361 | vLLM after 0.24.0 and before 0.31.0 contains a memory exhaustion vulnerability in MOSS-Audio model serving that caches a processor for every distinct request-controlled mmprocessorkwargs value in an unbounded dictionary. | From 0.24.0 before 0.31.0 | 0.31.0 | |
| CVE-2026-105922PYSEC-2026-4319 | A security flaw has been discovered in vllm-project vLLM up to 0.31.0. | Before 0.32.0 | 0.32.0 | |
| CVE-2026-105752GHSA-935w-9g4m-p28p, PYSEC-2026-4308 | vLLM: Harmony tool continuations drop `cache_salt`, restoring a cross-tenant prefix-cache membership oracleLow severity | Before 0.30.0 | 0.30.0 | |
| CVE-2026-105758GHSA-x6mc-67gf-chw4, PYSEC-2026-4310 | vLLM: Qwen2-VL / Qwen3-VL video samplers bound on request-controlled max_frames, which the num_frames ceiling does not reachModerate severity | From 0.24.0 before 0.30.0 | 0.30.0 | |
| CVE-2026-105760GHSA-58v5-2m8f-94pr, PYSEC-2026-4307 | vLLM: GLMGA video sampling permits request-driven CPU and memory exhaustionModerate severity | From 0.23.0rc2 before 0.30.0 | 0.30.0 | |
| CVE-2026-105754GHSA-ph72-cqr5-qpp7, PYSEC-2026-4309 | vLLM: Scale-out disaggregated multimodal transport trusts caller-supplied featuresModerate severity | Before 0.30.0 | 0.30.0 | |
| CVE-2026-105756GHSA-2823-qmq8-rwvj, PYSEC-2026-4306 | vLLM: Loose `cache_salt` validation lets a single request kill EngineCore on LMCache-MP deployments, uncaught downstream `ValueError` denial of serviceModerate severity | Before 0.30.0 | 0.30.0 | |
| CVE-2026-105759PYSEC-2026-4325 | vLLM is an inference and serving engine for large language models. | Before 0.30.0 | 0.30.0 | |
| CVE-2026-105757GHSA-85xf-c7hm-whqw, PYSEC-2026-4199 | vLLM: Structured-output request errors escape the request boundary and terminate the shared EngineCore, engine-fatal denial of service (3 sites)Moderate severity | Before 0.30.0 | 0.30.0 | |
| CVE-2026-105755GHSA-2phq-3phc-84px, PYSEC-2026-4198 | vLLM: Flash late-interaction scoring caches query embeddings under a caller-controlled request id, cross-request integrity break and induced errors on `/score` and `/rerank`Moderate severity | Before 0.30.0 | 0.30.0 | |
| CVE-2026-105753GHSA-ph3r-5jfg-f84f, PYSEC-2026-4197 | vLLM: Mirrored multimodal IPC caches desync after a rejected request, a later request reusing the same media hash trips a receiver assertion in the engine coreModerate severity | Before 0.28.0 | 0.28.0 | |
| CVE-2026-103241PYSEC-2026-4186 | A flaw has been found in vllm-project vLLM up to 0.26.0. | Before 0.27.0 | 0.27.0 | |
| CVE-2026-100653PYSEC-2026-4324 | vLLM is an inference and serving engine for large language models. | From 0.22.1 before 0.28.0 | 0.28.0 | |
| CVE-2026-100652PYSEC-2026-4185 | vLLM versions 0.22.0 through 0.23.0 fail to validate stoptokenids against vocabulary bounds in Rust HTTP and gRPC frontends, allowing out-of-vocabulary token IDs to reach MinTokensLogitsProcessor. | From 0.22.0 before 0.24.0 | 0.24.0 | |
| CVE-2026-100650PYSEC-2026-4184 | vLLM through 0.29.0 fetches and fully materializes remote or inline media before enforcing its documented media controls (the VLLMMAXAUDIOCLIPFILESIZEMB compressed-audio size cap, default 25 MB, and the per-modality --limit-mm-per-prompt... | Before 0.30.0 | 0.30.0 | |
| CVE-2026-94626PYSEC-2026-4008 | vLLM through 0.29.0 fails to validate the tpsize parameter in kvtransferparams on OpenAI-compatible completion endpoints, allowing attackers to allocate unbounded memory. | Before 0.30.0 | 0.30.0 | |
| CVE-2026-94625PYSEC-2026-4007 | vLLM through 0.29.0 contains a resource exhaustion vulnerability in MooncakeConnector where rejected prefill requests create ownerless transfer placeholders that are never reclaimed. | Before 0.30.0 | 0.30.0 | |
| CVE-2026-94624GHSA-frvg-qw88-hww9, PYSEC-2026-4006 | vllm p2p never connected zmq_exhaustionModerate severity | Before 0.32.0 | 0.32.0 | |
| CVE-2026-94623PYSEC-2026-4005 | vLLM through 0.29.0 contains a denial of service vulnerability in the NIXL connector's prefix caching implementation that fails to properly validate block counts across multi-prompt completion requests in prefill/decode disaggregated... | Before 0.30.0 | 0.30.0 | |
| CVE-2026-94622PYSEC-2026-4004 | vLLM versions through 0.29.0 contain a denial of service vulnerability in the NIXL connector's metadata handling for prefill/decode disaggregated deployments. | Before 0.30.0 | 0.30.0 | |
| CVE-2026-93989PYSEC-2026-4000 | vLLM through 0.29.0 fails to properly validate badwords token indices against the model's generation output width in SamplingParams.updatefromtokenizer(). | Before 0.30.0 | 0.30.0 | |
| CVE-2026-93841PYSEC-2026-3999 | vLLM through 0.29.0 contains a memory corruption vulnerability in the Triton bincountkernel where prompt token IDs index the penalty prompt-presence bitset without bounds checking against vocabulary size. | Before 0.30.0 | 0.30.0 | |
| CVE-2026-93840PYSEC-2026-3998 | vLLM before 0.29.0 validates allowedtokenids against tokenizer length instead of model output logits width in SamplingParams.validateallowedtokenids(). | Before 0.29.0 | 0.29.0 | |
| CVE-2026-93592PYSEC-2026-3997 | vLLM versions before 0.28.0 fail to validate the lower bound of token IDs in the /v1/embeddings and /pooling endpoints, allowing unauthenticated attackers to crash the engine by submitting negative token IDs. | Before 0.28.0 | 0.28.0 | |
| CVE-2026-93436PYSEC-2026-3996 | vLLM through 0.29.0 fails to properly clean up decode-side metadata for rejected inference requests in prefill/decode disaggregated deployments. | Before 0.30.0 | 0.30.0 | |
| CVE-2026-69147GHSA-8pw2-6jv3-mj5j, PYSEC-2026-4178 | vLLM: Request-selected PyNvVideoCodec GPU decode bypasses static VRAM reservationModerate severity | Before 0.28.0 | 0.28.0 | |
| CVE-2026-57173GHSA-hcwq-8wjf-3gcr, PYSEC-2026-4179 | vLLM: Unauthenticated audio decompression-bomb DoS in /v1/chat/completionsModerate severity | Before 0.24.0 | 0.24.0 | |
| CVE-2026-90553PYSEC-2026-3985 | vLLM before 0.28.0 contains a remote code execution vulnerability in the LlavaOnevision2 processor loader that ignores the trustremotecode parameter when loading remote processor classes. | Before 0.28.0 | 0.28.0 | |
| CVE-2026-73560GHSA-4hhp-h66f-j5j7, PYSEC-2026-3934 | vLLM: SSRF + arbitrary local file read in MiMoV2OmniMultiModalProcessor `_fetch_image` and audio loader bypass MediaConnector protectionsModerate severity | Before 0.26.0 | 0.26.0 | |
| CVE-2026-73558GHSA-7m6h-x95x-82q5, PYSEC-2026-3935 | vLLM: Cross-User Data Leak VulnerabilityModerate severity | Before 0.27.0 | 0.27.0 | |
| CVE-2026-73557GHSA-pr7f-p5mw-fc87, PYSEC-2026-3938 | vLLM: Incomplete CVE-2025-62164 remediation can be bypassed by concurrent prompt partsModerate severity | From 0.21.0 before 0.26.0 | 0.26.0 | |
| CVE-2026-73556GHSA-48jh-3gj7-fg8v, PYSEC-2026-3933 | vLLM: ReDoS via structured_outputs.regex in the lm-format-enforcer backend (no compile timeout), missed sibling of GHSA-rwxx-mrjm-wc2mModerate severity | Before 0.26.0 | 0.26.0 | |
| CVE-2026-73555GHSA-hwrm-c4cx-rf4j, PYSEC-2026-3937 | vLLM: Unauthenticated Internal Path and Username Disclosure via Validation Error MessagesModerate severity | Before 0.26.0 | 0.26.0 | |
| CVE-2026-71486GHSA-8737-qx52-hjff, PYSEC-2026-3936 | vLLM: Derender endpoints decode caller-supplied GenerateResponse token IDs without output boundsModerate severity | Before 0.26.0 | 0.26.0 | |
| CVE-2026-78684PYSEC-2026-4003 | vLLM before 0.27.0 fails to properly classify DeepStream as a GPU backend and omits pixel-limit enforcement in its decode path. | Before 0.27.0 | 0.27.0 | |
| CVE-2026-73559GHSA-87x5-vmc3-756j, PYSEC-2026-3704 | vLLM: Completion prompt lists fan out into unbounded engine requestsModerate severity | From 0.19.0 before 0.26.0 | 0.26.0 | |
| CVE-2026-54234GHSA-8wr5-jm2h-8r4f, PYSEC-2026-3542 | vLLM has Remote DoS via Invalid Recovered Token ReinjectionHigh severity | From 0.17.1 before 0.24.0 | 0.24.0 | |
| CVE-2026-55574GHSA-rwxx-mrjm-wc2m, PYSEC-2026-2304 | vLLM: ReDoS via structured_outputs.regex compiled without timeout in xgrammar and outlines backendsHigh severity | Before 0.24.0 | 0.24.0 | |
| CVE-2026-55514GHSA-33cg-gxv8-3p8g, PYSEC-2026-2303 | vLLM denial of service via prompt embeds on M-RoPE modelsHigh severity | From 0.12.0 before 0.24.0 | 0.24.0 | |
| CVE-2026-55646GHSA-v82g-2437-67m2, PYSEC-2026-2305 | vLLM: Speech-to-text upload size limit is enforced after full UploadFile readModerate severity | From 0.22.0 before 0.24.0 | 0.24.0 | |
| CVE-2026-54232PYSEC-2026-227 | vLLM is an inference and serving engine for large language models (LLMs). | Before 0.22.1 | 0.22.1 | |
| CVE-2026-54233GHSA-6pr9-rp53-2pmc, PYSEC-2026-3404 | vLLM: OOM Denial of Service via Audio Decompression BombModerate severity | Before 0.24.0 | 0.24.0 | |
| CVE-2026-54236GHSA-hgg8-fqqc-vfmw, PYSEC-2026-3408 | vLLM: incomplete CVE-2026-22778 fix leaks PIL repr addresses via Anthropic routerModerate severity | Before 0.24.0 | 0.24.0 | |
| CVE-2026-53923GHSA-5jv2-g5wq-cmr4, PYSEC-2026-3403 | vLLM: GGUF dequantize kernel int truncation exposes uninitialized GPU memory in multi-tenant servingModerate severity | From 0.5.5 before 0.24.0 | 0.24.0 | |
| CVE-2026-12491GHSA-8jr5-v98p-w75m, PYSEC-2026-3406 | vLLM: image EXIF Rotation & PNG tRNS Transparency Not Normalized, Causing Mismatch Between Model Input and ExpectationsModerate severity | From 0.11.0 before 0.24.0 | 0.24.0 | |
| CVE-2026-54235GHSA-7h4p-rffg-7823, PYSEC-2026-3405 | vLLM: temperature=NaN and temperature=Infinity bypass validation and propagate to GPU kernelsModerate severity | Before 0.24.0 | 0.24.0 | |
| CVE-2026-48746GHSA-94f4-hr76-p5j6, PYSEC-2026-226 | vLLM: OpenAI auth bypassCritical severity | From 0.3.0 before 0.22.0 | 0.22.0 | |
| CVE-2026-41523GHSA-q8gq-377p-jq3r, PYSEC-2026-2300 | vLLM: Security Check Bypass via assert Statement in Activation Function Loading Allows Arbitrary Code ExecutionHigh severity | Before 0.22.0 | 0.22.0 | |
| CVE-2026-5497GHSA-wcwg-c5fc-9vrc, PYSEC-2026-2302 | vLLM is vulnerable to an Out-of-Memory (OOM) Denial of Service (DoS) attack due to unbounded frame count processing in the `VideoMediaIO.load_base64()` methodHigh severity | From 0.8.0 before 0.19.0 | 0.19.0 | |
| CVE-2026-47155GHSA-3ww4-5jv9-j5gm, PYSEC-2026-2301 | vLLM's Artifact Pin Decay allows pinned deployments to load unpinned code, weights, and processorsModerate severity | Before 0.22.0 | 0.22.0 | |
| CVE-2026-9540GHSA-98f3-hwg4-4rf7, PYSEC-2026-3407 | vllm has Improper Resource Shutdown or ReleaseModerate severity | Up to and including 0.19.0 | No fixed version listed | |
| CVE-2026-44223GHSA-83vm-p52w-f9pw, PYSEC-2026-145 | vLLM: extract_hidden_states speculative decoding crashes server on any request with penalty parametersModerate severity | From 0.18.0 before 0.20.0 | 0.20.0 | |
| CVE-2026-44222GHSA-hpv8-x276-m59f, PYSEC-2026-3409 | vLLM Vulnerable to Remote DoS via Special-Token PlaceholdersModerate severity | From 0.6.1 before 0.20.0 | 0.20.0 | |
| CVE-2026-7141GHSA-x368-4g9h-fvv4, PYSEC-2026-2306 | vLLM makes Use of Uninitialized ResourceLow severity | Before 0.19.1 | 0.19.1 | |
| CVE-2026-34755GHSA-pq5c-rjhq-qp7p, PYSEC-2026-144 | vLLM: Denial of Service via Unbounded Frame Count in video/jpeg Base64 ProcessingModerate severity | From 0.7.0 before 0.19.0 | 0.19.0 | |
| CVE-2026-34753GHSA-pf3h-qjgv-vcpr, PYSEC-2026-3410 | vLLM: Server-Side Request Forgery (SSRF) in `download_bytes_from_url `Moderate severity | From 0.16.0 before 0.19.0 | 0.19.0 | |
| CVE-2026-34756GHSA-3mwp-wvh9-7528, PYSEC-2026-2298 | vLLM: Unauthenticated OOM Denial of Service via Unbounded `n` Parameter in OpenAI API ServerModerate severity | From 0.1.0 before 0.19.0 | 0.19.0 | |
| CVE-2026-34760GHSA-6c4r-fmh3-7rh8, PYSEC-2026-2299 | vLLM: Processing differential in multi-channel audio downmixing enables hidden-input/moderation bypass for audio modelsModerate severity | From 0.5.5 before 0.18.0 | 0.18.0 | |
| CVE-2026-27893GHSA-7972-pg2x-xr59, PYSEC-2026-2297 | vLLM has Hardcoded Trust Override in Model Files Enables RCE Despite Explicit User Opt-OutHigh severity | From 0.10.1 before 0.18.0 | 0.18.0 | |
| CVE-2026-25960GHSA-v359-jj2v-j536, PYSEC-2026-3411 | vLLM has SSRF Protection BypassModerate severity | From 0.15.1 before 0.17.0 | 0.17.0 | |
| CVE-2026-22778GHSA-4r2x-xpjr-7cvv, PYSEC-2026-565 | vLLM has RCE In Video ProcessingCritical severity | From 0.8.3 before 0.14.1 | 0.14.1 | |
| CVE-2026-24779GHSA-qh4c-xf7m-gxfc, PYSEC-2026-2020 | vLLM vulnerable to Server-Side Request Forgery (SSRF) through MediaConnectorHigh severity | Before 0.14.1 | 0.14.1 | |
| CVE-2026-22807GHSA-2pc9-4j83-qjmr, PYSEC-2026-2010 | vLLM affected by RCE via auto_map dynamic module loading during model initializationHigh severity | From 0.10.1 before 0.14.0 | 0.14.0 | |
| CVE-2026-22773GHSA-grg2-63fw-f2qr, PYSEC-2026-143 | vLLM is vulnerable to DoS in Idefics3 vision models via image payload with ambiguous dimensionsModerate severity | From 0.6.4 before 0.12.0 | 0.12.0 | |
| CVE-2026-56340GHSA-mcmc-2m55-j8jj, PYSEC-2026-250 | vLLM introduced enhanced protection for CVE-2025-62164High severity | From 0.10.2 before 0.13.0 | 0.13.0 | |
| CVE-2025-66448GHSA-8fr4-5q9j-m8gm, PYSEC-2026-2015 | vLLM vulnerable to remote code execution via transformers_utils/get_configHigh severity | Before 0.11.1 | 0.11.1 | |
| CVE-2025-62426GHSA-69j4-grxj-j64p, PYSEC-2026-2012 | vLLM vulnerable to DoS via large Chat Completion or Tokenization requests with specially crafted `chat_template_kwargs`Moderate severity | From 0.5.5 before 0.11.1 | 0.11.1 | |
| CVE-2025-62372GHSA-pmqf-x6x8-p7qw, PYSEC-2026-2019 | vLLM vulnerable to DoS with incorrect shape of multimodal embedding inputsHigh severity | From 0.5.5 before 0.11.1 | 0.11.1 | |
| CVE-2025-62164GHSA-mrw7-hf4f-83pf, PYSEC-2026-2018 | vLLM deserialization vulnerability leading to DoS and potential RCEHigh severity | From 0.10.2 before 0.11.1 | 0.11.1 | |
| CVE-2025-6242GHSA-3f6c-7fw2-ppm4, PYSEC-2026-2011 | vLLM is vulnerable to Server-Side Request Forgery (SSRF) through `MediaConnector` classHigh severity | From 0.5.0 before 0.11.0 | 0.11.0 | |
| CVE-2025-61620GHSA-6fvq-23cw-5628, PYSEC-2026-2013 | vLLM: Resource-Exhaustion (DoS) through Malicious Jinja Template in OpenAI-Compatible ServerModerate severity | From 0.5.1 before 0.11.0 | 0.11.0 | |
| CVE-2025-59425GHSA-wr9h-g72x-mwhm, PYSEC-2026-2026 | vLLM is vulnerable to timing attack at bearer authHigh severity | Before 0.11.0 | 0.11.0 | |
| CVE-2025-9141GHSA-79j6-g2m3-jgfw, PYSEC-2026-2014 | vLLM has remote code execution vulnerability in the tool call parser for Qwen3-CoderHigh severity | From 0.10.0 before 0.10.1.1 | 0.10.1.1 | |
| CVE-2025-48956GHSA-rxc4-3w6r-4v47, PYSEC-2026-2021 | vllm API endpoints vulnerable to Denial of Service AttacksHigh severity | From 0.1.0 before 0.10.1.1 | 0.10.1.1 | |
| CVE-2025-48944GHSA-vrq3-r879-7m65, PYSEC-2026-2023 | vLLM Tool Schema allows DoS via Malformed pattern and type FieldsModerate severity | From 0.8.0 before 0.9.0 | 0.9.0 | |
| CVE-2025-48943GHSA-9hcf-v7m4-6m2j, PYSEC-2025-55 | vLLM allows clients to crash the openai server with invalid regexModerate severity | From 0.8.0 before 0.9.0 | 0.9.0 | |
| CVE-2025-48942GHSA-6qc9-v4r8-22xg, PYSEC-2025-54 | vLLM DOS: Remotely kill vllm over http with invalid JSON schemaModerate severity | From 0.8.0 before 0.9.0 | 0.9.0 | |
| CVE-2025-46722GHSA-c65p-x677-fgj6, PYSEC-2025-43 | vLLM has a Weakness in MultiModalHasher Image Hashing ImplementationModerate severity | Before 0.9.0 | 0.9.0 | |
| CVE-2025-46570GHSA-4qjh-9fv9-r85r, PYSEC-2025-53 | Potential Timing Side-Channel Vulnerability in vLLM’s Chunk-Based Prefix CachingLow severity | Before 0.9.0 | 0.9.0 | |
| CVE-2025-71379GHSA-j828-28rj-hfhp | vLLM vulnerable to Regular Expression Denial of ServiceModerate severity | From 0.6.3 before 0.9.0 | 0.9.0 | |
| CVE-2025-48887GHSA-w6q7-j642-7c25, PYSEC-2025-50 | vLLM has a Regular Expression Denial of Service (ReDoS, Exponential Complexity) Vulnerability in `pythonic_tool_parser.py`Moderate severity | From 0.6.4 before 0.9.0 | 0.9.0 | |
| CVE-2025-47277GHSA-hjq4-87xh-g4fv, PYSEC-2026-567 | vLLM Allows Remote Code Execution via PyNcclPipe Communication ServiceCritical severity | From 0.6.5 before 0.8.5 | 0.8.5 | |
| CVE-2025-30165GHSA-9pcc-gvx5-r5wm, PYSEC-2026-2017 | Remote Code Execution Vulnerability in vLLM Multi-Node Cluster ConfigurationHigh severity | From 0.5.2 before 0.10.0 | 0.10.0 | |
| CVE-2025-46560GHSA-vc6m-hm49-g9qg, PYSEC-2026-2022 | vLLM: Quadratic Time Complexity in Input Token Processing leads to denial of serviceModerate severity | From 0.8.0 before 0.8.5 | 0.8.5 |
Other PyPI packages
- django162 advisories
- open-webui161 advisories
- praisonai105 advisories
- plone101 advisories