CVE-2026-71486

Summary

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 structures are processed by OnlineDerenderer and tokenizer.decode before max_model_len, max_tokens, max_num_seqs, or response-size limits are enforced, allowing an authenticated API client to consume excessive CPU and memory and produce oversized responses. This issue is fixed in version 0.26.0.

Affected Software

VendorProductVersion RangeStatus
vllm-projectvllm< 0.26.0affected

Weaknesses

  • CWE-400: CWE-400: Uncontrolled Resource Consumption
  • CWE-770: CWE-770: Allocation of Resources Without Limits or Throttling

References