Allocation of Resources Without Limits or Throttling Affecting vllm package, versions [0.22.0,0.24.0)
Threat Intelligence
The probability is the direct output of the EPSS model, and conveys an overall sense of the threat of exploitation in the wild. The percentile measures the EPSS probability relative to all known EPSS scores. Note: This data is updated daily, relying on the latest available EPSS model version. Check out the EPSS documentation for more details.
Do your applications use this vulnerable package?
In a few clicks we can analyze your entire application and see what components are vulnerable in your application, and suggest you quick fixes.
Test your applications- Snyk IDSNYK-PYTHON-VLLM-21707829
- published25 Oct 2026
- disclosed25 Oct 2026
- creditRex Liu
Introduced: 25 Oct 2026
NewCVE-2026-116173 (opens in a new tab)How to fix?
Upgrade vllm to version 0.24.0 or higher.
Overview
vllm is an A high-throughput and memory-efficient inference and serving engine for LLMs
Affected versions of this package are vulnerable to Allocation of Resources Without Limits or Throttling via the Rust gRPC frontend's handling of output_candidates.token_ids in Generate and GenerateStream requests. The Rust gRPC converter maps ResponseOptions.output_candidates.token_ids directly into SamplingParams.logprob_token_ids without enforcing the MAX_LOGPROB_TOKEN_IDS = 128 limit, vocabulary range checks, or the logprobs == len(logprob_token_ids) consistency requirement that the Python SamplingParams validator applies. An attacker can send a generation request with output_logprobs enabled and more than 128 output_candidates.token_ids, which bypasses the northbound validator, reaches the V1 GPU logprob sampler's fixed-width (max_num_reqs, MAX_LOGPROB_TOKEN_IDS) allocation, and triggers a fatal ValueError that crashes EngineCore.