CVE search
Every published CVE plus open-source advisories (NVD, OSV, GitHub), with CVSS, EPSS and whether it is being exploited in the wild (CISA KEV). Updated every few hours. Our scans match these against the technologies they detect.
Every published CVE plus open-source advisories (NVD, OSV, GitHub), with CVSS, EPSS and whether it is being exploited in the wild (CISA KEV). Updated every few hours. Our scans match these against the technologies they detect.
| 21/09/2026 |
| CVE-2026-93989 | 보통4.3 | vLLM through 0.29.0 fails to properly validate bad_words token indices against the model's generation output width in SamplingParams.update_from_tokenizer(). Attackers can supply out-of-bounds token indices that corrupt logits memory of concurrent requests, causing different in-flight HTTP requests to return incorrect | 19/09/2026 |
| CVE-2026-93841 | 보통5.3 | vLLM through 0.29.0 contains a memory corruption vulnerability in the Triton _bincount_kernel where prompt token IDs index the penalty prompt-presence bitset without bounds checking against vocabulary size. Attackers can submit multimodal audio requests with tokens equal to vocabulary size, causing out-of-bounds writes | 18/09/2026 |
| CVE-2026-93840 | 보통5.3 | vLLM before 0.29.0 validates allowed_token_ids against tokenizer length instead of model output logits width in SamplingParams._validate_allowed_token_ids(). Attackers can supply token IDs above the output vocabulary that pass validation, causing LogitBiasState to corrupt GPU logits state and allow concurrent requests | 18/09/2026 |
| CVE-2026-93592 | 높음7.5 | 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 a CUDA device-side assertion that poisons the GPU context, c | 18/09/2026 |
| CVE-2026-93436 | 높음7.5 | 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 worker restarts. | 17/09/2026 |
| CVE-2026-69147 | 보통6.5 | vLLM is an inference and serving engine for large language models. Prior to 0.28.0, request bodies for Chat Completions and Responses can set media_io_kwargs.video.video_backend to pynvvideocodec, and MediaConnector.fetch_video forwards that choice to VideoMediaIO even when startup configuration selected a software dec | 16/09/2026 |
| CVE-2026-57173 | 보통6.5 | vLLM is an inference and serving engine for large language models. Prior to 0.24.0, the input_audio handling path for /v1/chat/completions calls AudioMediaIO.load_bytes or AudioMediaIO.load_file without passing VLLM_MAX_AUDIO_DECODE_DURATION_S to the shared audio decoder. An unauthenticated client can therefore submit | 16/09/2026 |
| CVE-2026-90555 | 보통6.5 | vLLM versions before 0.28.0 fail to validate audio sample rate headers in the transcription endpoint, allowing authenticated clients to bypass duration checks. Attackers can submit forged FLAC headers with inflated sample rates to trigger excessive memory allocation and crash the API server process affecting all tenant | 12/09/2026 |
| CVE-2026-90554 | 높음7.5 | 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 max_duration_s or max_decode_bytes parameters, so neither V | 12/09/2026 |
| CVE-2026-90553 | 높음7.8 | 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 processing_llava_onevision2.py that executes with vLLM process author | 12/09/2026 |
| CVE-2026-37237 | 높음7.5 | vLLM up to and including 0.17.0 allows remote attackers to cause a Denial of Service via memory exhaustion. The AsyncMediaIO.fetch_audio and AsyncMediaIO.fetch_image functions in multimodal/inputs.py fetch user-supplied media URLs using aiohttp and call r.read() without enforcing a maximum response size, allowing an at | 28/08/2026 |
| CVE-2026-78684 | 보통5.3 | vLLM before 0.27.0 fails to properly classify DeepStream as a GPU backend and omits pixel-limit enforcement in its decode path. Unauthenticated attackers can activate DeepStream at request time to initialize the process-wide GPU decode pool and submit video that bypasses resource controls, causing partial denial of ser | 25/08/2026 |
| CVE-2026-73560 | 보통6.5 | vLLM is an inference and serving engine for large language models. Prior to 0.26.0, the MiMoV2OmniMultiModalProcessor in vllm/transformers_utils/processors/mimo_v2_omni.py passes attacker-controlled image and audio strings through _fetch_image, requests.get, and Image.open instead of MediaConnector, bypassing allowed_m | 17/08/2026 |
| CVE-2026-71486 | 보통4.3 | vLLM is an inference and serving engine for large language models. Prior to 0.26.0, the /v1/completions/derender and /v1/chat/completions/derender endpoints accept caller-supplied GenerateResponse objects whose generate_responses, choices, token_ids, prompt_logprobs, logprobs.content, top_logprobs, and routed_experts s | 17/08/2026 |
| CVE-2026-73559 | 보통6.5 | vLLM is an inference and serving engine for large language models. From 0.19.0 until 0.26.0, the /v1/completions CompletionRequest.prompt field in vllm/entrypoints/openai/completion/protocol.py accepts an unbounded list[str] or list[list[int]], prompt_to_seq() in vllm/renderers/inputs/preprocess.py and OnlineRenderer.p | 13/08/2026 |
| CVE-2026-73558 | 보통5.3 | vLLM is an inference and serving engine for large language models. Prior to 0.27.0, an integer overflow in blockIdx.x * 2 * d in activation_kernels.cu can cause act_and_mul_kernel to consume another batched user's input, allowing a request processed in the same inference batch to receive a partial or complete copy of a | 13/08/2026 |
| CVE-2026-73557 | 보통5.3 | vLLM is an inference and serving engine for large language models. From 0.20.2rc0 until 0.26.0, safe_load_prompt_embeds in vllm/renderers/embed_utils.py uses torch.sparse.check_sparse_tensor_invariants, whose process-global save, enable, and restore state can be raced by concurrent prompt_embeds parts submitted to POST | 13/08/2026 |