vLLM vulnerabilities
56 known vulnerabilities in vLLM from Vllm-project. 56 are rated high or critical. Each entry lists the affected versions and the release that fixes it, where the vendor has published one.
Latest first
- 7.5 High severity
vllm: Uncontrolled Resource Consumption
vLLM from 0.10.1 before 0.31.0 forwards a request-controlled mm_processor_kwargs.max_patches value to the Hugging Face processor for Cohere2 Vision models without a server-side upper bound, causing quadratic CPU work during multimodal preprocessing. Remote cli
- 9.8 Critical severity
vLLM is an inference and serving engine for large language models.
vLLM is an inference and serving engine for large language models. Prior to 0.32.0, ImageMediaIO passes attacker-controlled image bytes to Pillow through Image.open(BytesIO(data)) without restricting decoder selection, and the MIME label does not constrain the
- 8.8 High severity
vllm: Deserialization of Untrusted Data
vLLM is an inference and serving engine for large language models. Prior to 0.32.0, an authenticated caller can place forged shared-memory address and monotonic_id values in public multimodal features.kwargs_data. In scale-out deployments with shared-memory mu
- 7.7 High severity
vllm: Use of Uninitialized Resource
vLLM from 0.8.5 before 0.31.0 has a denial of service flaw. LogprobsTensors.empty_cpu() in vllm/v1/outputs.py allocates an uninitialized int32 tensor, and the V1 model runner leaves prompt positions reported as computed by a KV connector unwritten. A client sh
- 7.5 High severity
vllm: Allocation of Resources Without Limits or Throttling
vLLM from 0.18.0 before 0.31.0 contains a resource exhaustion vulnerability that allows API clients to bypass the video/jpeg frame-count cap by overriding num_frames through media_io_kwargs. Attackers can set num_frames to -1 with a large inline video so Video
- 7.5 High severity
vLLM from 0.20.0 before 0.30.0 contains a denial of service vulnerability that allows remote…
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. Attackers can replay a valid multimodal
- 7.5 High severity
vllm: Uncontrolled Resource Consumption
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.MAX_IMAGE_PIXELS to None. Attackers can submit a pathologic
- 8.2 High severity
vllm: Server-Side Request Forgery
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. Attackers can make a P2pNcclConnector prefill worker ZMQ-connect
- 7.5 High severity
vLLM from 0.22.0 before 0.24.0 contains an improper input validation vulnerability in the Rust…
vLLM from 0.22.0 before 0.24.0 contains an improper input validation vulnerability in the Rust gRPC Generate and GenerateStream service that forwards unchecked prompt token IDs to EngineCore. Attackers can submit requests with out-of-vocabulary prompt.token_id
- 7.5 High severity
vLLM from 0.22.0 before 0.24.0 contains a denial of service vulnerability that allows remote…
vLLM from 0.22.0 before 0.24.0 contains a denial of service vulnerability that allows remote attackers to terminate EngineCore by bypassing logprob_token_ids limits in the Rust gRPC frontend. Attackers can send Generate or GenerateStream requests with output_l
- 7.5 High severity
vLLM from 0.22.0 before 0.24.0 contains an improper input validation vulnerability in the Rust…
vLLM from 0.22.0 before 0.24.0 contains an improper input validation vulnerability in the Rust gRPC converter that allows remote attackers to bypass max_logprobs limits via CandidateTokens.top_n. Attackers can send Generate or GenerateStream requests with over
- 7.5 High severity
vllm: Uncontrolled Resource Consumption
vLLM from 0.7.3 before 0.31.0 contains an uncontrolled resource consumption vulnerability that allows remote clients to control multimodal processor parameters by supplying per-request mm_processor_kwargs. Attackers can set large image size, patch count, crop
- 8.1 High severity
vllm: Improper Control of Generation of Code
vLLM from 0.7.3 before 0.31.0 forwards the request-supplied mm_processor_kwargs.code_revision to Transformers AutoProcessor.from_pretrained, letting clients select which revision of a model's dynamic processor code is imported when trust_remote_code is enabled
- 7.5 High severity
vllm: Allocation of Resources Without Limits or Throttling
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 mm_processor_kwargs value in an unbounded dictionary. Unauthenticated attackers can send di
- 7.5 High severity
vllm: Uncaught Exception
vLLM before 0.31.0 skips the min_tokens <= max_tokens check for /inference/v1/generate requests with a null max_tokens, so the EngineCore input thread later dies on an uncaught VLLMValidationError while /health still returns 200. Unauthenticated remote attacke
- 7.5 High severity
vllm: Insufficient Verification of Data Authenticity
vLLM before 0.31.0 uses a client-supplied multimodal uuid verbatim as cache identity with no consistency check on cache hits, allowing engine crashes. Unauthenticated attackers can desynchronize the P0/P1 processor caches to trigger a fatal token-placeholder m
- 7.5 High severity
vllm: Uncontrolled Resource Consumption
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, the Rust frontend's track_http_metrics middleware records the raw HTTP method token as a Prometheus label for requests reaching registered routes. An unauthenticated attacker c
- 7.5 High severity
vLLM: Improper Resource Shutdown or Release
A flaw has been found in vllm-project vLLM up to 0.26.0. This vulnerability affects unknown code of the file rust/src/parser/src/unified/gemma4.rs of the component Gemma4UnifiedParser. Executing a manipulation can lead to denial of service. The attack may be l
- 7.5 High severity
vLLM versions 0.22.0 through 0.23.0 fail to validate stop_token_ids against vocabulary bounds…
vLLM versions 0.22.0 through 0.23.0 fail to validate stop_token_ids against vocabulary bounds in Rust HTTP and gRPC frontends, allowing out-of-vocabulary token IDs to reach MinTokensLogitsProcessor. Attackers can submit requests with min_tokens greater than ze
- 7.5 High severity
vLLM versions before 0.29.0 contain a denial-of-service vulnerability in the cache_salt…
vLLM versions before 0.29.0 contain a denial-of-service vulnerability in the cache_salt parameter accepted on OpenAI-compatible and Anthropic API endpoints, which lacks maximum length validation and is processed on the single EngineCore scheduler thread. Unaut
- 7.5 High severity
vllm: Missing Release of Memory after Effective Lifetime
vLLM Mooncake connector through 0.29.0 fails to properly manage GPU KV cache block ownership when concurrent child requests share a single transfer ID in prefill/decode disaggregated deployments. Attackers can trigger GPU memory exhaustion by submitting comple
- 7.5 High severity
vllm: Memory Allocation with Excessive Size Value
vLLM through 0.29.0 fails to validate the tp_size parameter in kv_transfer_params on OpenAI-compatible completion endpoints, allowing attackers to allocate unbounded memory. Attackers can supply arbitrary tp_size values in prefill/decode disaggregated deployme
- 7.5 High severity
vllm: Allocation of Resources Without Limits or Throttling
vLLM through 0.29.0 contains a denial of service vulnerability in P2P KV offloading when OffloadingConnector is configured with TieringOffloadingSpec and a peer-to-peer secondary tier. Attackers can supply arbitrary remote host and port values in kv_transfer_p
- 7.5 High severity
vllm: Reachable Assertion
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 deployments. Attackers can
- 7.5 High severity
vllm: Uncaught Exception
vLLM versions through 0.29.0 contain a denial of service vulnerability in the NIXL connector's metadata handling for prefill/decode disaggregated deployments. Attackers can send requests with incomplete kv_transfer_params dictionary entries to trigger an uncau
- 7.5 High severity
vllm: Improper Validation of Array Index
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. A single request with a negative token ID triggers
- 7.5 High severity
vllm: Missing Release of Memory after Effective Lifetime
vLLM through 0.29.0 fails to properly clean up decode-side metadata for rejected inference requests in prefill/decode disaggregated deployments. Remote attackers can submit requests with max_tokens=0 to exhaust decode-worker memory without bound until the work
- 7.5 High severity
vLLM: Uncontrolled Resource Consumption
vLLM versions >=0.10.2 and <0.28.0 do not apply any audio decode-size or duration limit when extracting audio from video input for NanoNemotronVL models. In nano_nemotron_vl.py, _extract_audio_from_videos calls load_audio_pyav(BytesIO(video_bytes)) without the
- 7.8 High severity
vLLM: Improper Control of Generation of Code
vLLM before 0.28.0 contains a remote code execution vulnerability in the LlavaOnevision2 processor loader that ignores the trust_remote_code parameter when loading remote processor classes. Attackers can craft a malicious model with arbitrary code in processin
- 7.5 High severity
vllm: Inefficient Regular Expression Complexity
vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Prior to 0.24.0, the structured_outputs.regex API parameter passes a user-supplied regular expression string directly to the grammar compiler backends with no compilation tim
- 7.5 High severity
vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs.
vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Prior to 0.24.0, a frontend-legal multi-request speculative decoding workload can cause the rejection sampler to produce a recovered token equal to the model vocabulary size
- 8.8 High severity
vllm: Uncontrolled Search Path Element
vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.22.1, the vLLM Dockerfile is vulnerable to a dependency confusion attack through the flashinfer-jit-cache package. The package is installed from a custom index (flashinfer.ai/
- 7.5 High severity
vllm: Exposure of Sensitive Information to an Unauthorized Actor
vLLM is an inference and serving engine for large language models (LLMs). From 0.5.5 until 0.23.1rc0, integer truncation of tensor dimensions in vLLM's GGUF dequantize kernels (csrc/quantization/gguf/gguf_kernel.cu) causes partial tensor processing. The output
- 9.1 Critical severity
vllm: Inconsistent Interpretation of HTTP Requests
vLLM is an inference and serving engine for large language models (LLMs). From 0.3.0 until 0.22.0, a vulnerability in ASGI web servers and starlette's trust on those web servers enables an authentication bypass of the OpenAI API AuthenticationMiddleware. It al
- 7.5 High severity
vllm: Improper Control of Generation of Code
vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.22.0, an assert-based security check in vLLM's activation function loading allows any unauthenticated attacker to achieve arbitrary code execution on the server by publishing
- 7.5 High severity
vllm: Improper Validation of Array Index
vLLM is an inference and serving engine for large language models (LLMs). From 0.6.1 to before 0.20.0, there is a a Token Injection vulnerability in vLLM’s multimodal processing. Unauthenticated, text-only prompts that spell special tokens are interpreted as c
- 7.1 High severity
vLLM is an inference and serving engine for large language models (LLMs).
vLLM is an inference and serving engine for large language models (LLMs). From version 0.5.5 to before version 0.18.0, Librosa defaults to using numpy.mean for mono downmixing (to_mono), while the international standard ITU-R BS.775-4 specifies a weighted down
- 8.8 High severity
vllm: Protection Mechanism Failure
vLLM is an inference and serving engine for large language models (LLMs). Starting in version 0.10.1 and prior to version 0.18.0, two model implementation files hardcode `trust_remote_code=True` when loading sub-components, bypassing the user's explicit `--tru
- 9.8 Critical severity
vllm: Server-Side Request Forgery
vLLM is an inference and serving engine for large language models (LLMs). The SSRF protection fix for CVE-2026-24779 add in 0.15.1 can be bypassed in the load_from_url_async method due to inconsistent URL parsing behavior between the validation layer and the a
- 9.8 Critical severity
vllm: Insertion of Sensitive Information into Log File
vLLM is an inference and serving engine for large language models (LLMs). From 0.8.3 to before 0.14.1, when an invalid image is sent to vLLM's multimodal endpoint, PIL throws an error. vLLM returns this error to the client, leaking a heap address. With this le