Deserialization in Modeltc Lightllm
CVE-2026-111160
LightLLM through 1.2.0 contains a denial of service vulnerability that allows unauthenticated attackers to crash the inference worker by supplying the internal pd_kv_trans_params field to the /generate endpoint. Attackers can send a single request whose parameters cause pickle.loads to fail in InferSamplingParams, killing the inference thread while /health stays up and all later requests hang until restart.
Vulnerability class: Insecure Deserialization
Published · last modified .
CVSS v3 metric
CVSS v3 base score 7.5 (High). Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H.
EPSS exploit prediction
No EPSS score from FIRST.org yet. New CVEs usually get one within a few days; no score means no signal yet, not low risk. How EPSS works.
Affected products
Weakness classification (CWE)
References
- github.com/ModelTC/LightLLM/issues/1629 (issue-tracking, exploit)
- github.com/ModelTC/lightllm/blob/v1.2.0/lightllm/server/api_lightllm.py (technical-description)
- github.com/ModelTC/lightllm/blob/v1.2.0/lightllm/server/core/objs/sampling_para… (technical-description)
- github.com/ModelTC/lightllm/blob/v1.2.0/lightllm/server/router/model_infer/infe… (technical-description)
- github.com/ModelTC/LightLLM (product)
- www.vulncheck.com/advisories/lightllm-through-1.2.0-unauthenticated-dos-via-pd-… (third-party-advisory)
Frequently asked questions
- What is CVE-2026-111160?
- CVE-2026-111160 is a high-severity vulnerability in Modeltc Lightllm, classified under Deserialization of Untrusted Data. CVSS score: 7.5/10. Published 2026-10-15.
- How severe is CVE-2026-111160?
- High severity. CVSS v3 base score is 7.5 out of 10.