← Vulnerability feed

Vulnerability record · CVE-2026-100651 · published 26 September 2026

CVE-2026-100651: Uncontrolled resource consumption vulnerability

vLLM before 0.29.0 fails to enforce decoder prompt-length validation on the disaggregated serving endpoint /inference/v1/generate. When the request contains a 'features' (multimodal) payload, vllm/entrypoints/serve/disagg/serving.py builds a multimodal EngineInput directly from the caller-supplied token_ids, and GenerateRequest.token_ids (vllm/entrypoints/serve/disagg/protocol.py) is not checked against model_config.max_model_len. For multimodal processors that report skip_prompt_length_check=True (for example Nemotron Parse, Whisper, and FireRedLID), InputProcessor._validate_prompt_len() returns immediately for both encoder and decoder prompts, so an overlong prompt becomes an EngineCoreRequest and reaches the worker input-batch copy into a fixed max_model_len-wide NumPy row. A client able to reach the endpoint on an affected model configuration can therefore submit an overlong token_ids list to trigger a worker failure and denial of service. Fixed in 0.29.0.

7.1 CVSS 4.0 High CWE-400 · Uncontrolled resource consumption Received
7.1CVSS 4.0 base score
—EPSS exploitation probability, 30 days
NoNot in CISA KEV
0Affected product versions listed by NVD
2References
26 Sep 2026Last modified by NVD

Description

vLLM before 0.29.0 fails to enforce decoder prompt-length validation on the disaggregated serving endpoint /inference/v1/generate. When the request contains a 'features' (multimodal) payload, vllm/entrypoints/serve/disagg/serving.py builds a multimodal EngineInput directly from the caller-supplied token_ids, and GenerateRequest.token_ids (vllm/entrypoints/serve/disagg/protocol.py) is not checked against model_config.max_model_len. For multimodal processors that report skip_prompt_length_check=True (for example Nemotron Parse, Whisper, and FireRedLID), InputProcessor._validate_prompt_len() returns immediately for both encoder and decoder prompts, so an overlong prompt becomes an EngineCoreRequest and reaches the worker input-batch copy into a fixed max_model_len-wide NumPy row. A client able to reach the endpoint on an affected model configuration can therefore submit an overlong token_ids list to trigger a worker failure and denial of service. Fixed in 0.29.0.

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:

References

LinkTags
https://github.com/vllm-project/vllm/security/advisories/GHSA-3mqx-f33v-vgp9
https://www.vulncheck.com/advisories/vllm-before-0.29.0-denial-of-service-via-decoder-prompt-length-bypass

Track CVE-2026-100651 inside VULONE

Watch it alongside the ransomware crews, C2 infrastructure and forum chatter that reference it, query it through the API and pull it into your SIEM over TAXII.

Start free Open in platform

Related vulnerabilities

Same products first, then exploited flaws of the same weakness class.

7.5CVE-2026-28318SolarWinds Serv-U unauthenticated POST request denial of serviceSolarWinds Serv-U crashes when it receives a specially crafted POST request using Content-Encoding: deflate, and the crash occurs without authenticat…KEVEPSS 1.9%analysed7.5CVE-2026-45498Microsoft Defender antimalware platform uncontrolled resource consumption DoSCVE-2026-45498 is a denial of service flaw in the Microsoft Defender antimalware platform, classified as uncontrolled resource consumption (CWE-400).…KEVEPSS 1.3%analysed7.5CVE-2023-44487HTTP/2 Rapid Reset stream cancellation denial of serviceThe HTTP/2 protocol permits a client to cancel many streams quickly, and the server's handling of those resets consumes disproportionate resources. T…KEVEPSS 100%analysed7.5CVE-2023-38180Microsoft .NET and Visual Studio uncontrolled resource consumption DoSCVE-2023-38180 is a denial-of-service flaw in Microsoft .NET, ASP.NET Core and Visual Studio 2022, classified as uncontrolled resource consumption (C…KEVEPSS 14%analysed5.9CVE-2004-1464Cisco IOS Telnet VTY resource exhaustion denial of serviceCisco IOS 12.2(15) and earlier fails to properly handle crafted TCP connections to the Telnet or reverse Telnet port, exhausting virtual terminal (VT…KEVEPSS 4.8%analysed10.0CVE-2021-44228Apache Log4j2 JNDI lookup remote code executionApache Log4j2 versions 2.0-beta9 through 2.15.0 (excluding 2.12.2, 2.12.3, and 2.3.1) do not protect against attacker-controlled LDAP and other JNDI …KEVEPSS 100%analysed8.6CVE-2020-3569Cisco IOS XR DVMRP IGMP packet handling memory exhaustionCisco IOS XR Software mishandles IGMP packets in its Distance Vector Multicast Routing Protocol (DVMRP) feature, allowing crafted traffic to crash th…KEVEPSS 3.3%analysed8.6CVE-2020-3566Cisco IOS XR DVMRP IGMP queue flaw causes memory exhaustionCisco IOS XR Software mishandles queue management for IGMP packets in its DVMRP feature, allowing uncontrolled memory consumption. An unauthenticated…KEVEPSS 3.7%analysed

Source: NIST National Vulnerability Database (record CVE-2026-100651), CISA KEV, FIRST EPSS. This page is refreshed as NVD updates the record.