CVE-2026-22773
2 advisories.
PyPI/vllm CVE-2026-22773: medium 6.5
vLLM is vulnerable to DoS in Idefics3 vision models via image payload with ambiguous dimensions
Action: Upgrade vllm to 0.12.0 or later.
pip install vllm==0.12.0Affected
| introduced | fixed | last affected |
|---|---|---|
| 0.6.4 | 0.12.0 | - |
Fixed versions
| 0.12.0 |
Sources
- osv: GHSA-grg2-63fw-f2qr
- https://github.com/vllm-project/vllm/security/advisories/GHSA-grg2-63fw-f2qr
- https://nvd.nist.gov/vuln/detail/CVE-2026-22773
- https://github.com/vllm-project/vllm/pull/29881
- https://github.com/vllm-project/vllm/commit/0ec84221718d920c3f46da879cc354f94b8fb59e
- https://github.com/pypa/advisory-database/tree/main/vulns/vllm/PYSEC-2026-143.yaml
- https://github.com/vllm-project/vllm
CVE-2026-22773 · published · modified · JSON
PyPI/vllm CVE-2026-22773: high 7.5
vLLM is an inference and serving engine for large language models (LLMs). In versions from 0.6.4 to before 0.12.0, users can crash the vLLM engine serving multimodal models that use the Idefics3 visio
Action: Upgrade vllm to 0.12.0 or later.
pip install vllm==0.12.0Affected
| introduced | fixed | last affected |
|---|---|---|
| 0.6.4 | 0.12.0 | - |
Fixed versions
| 0.12.0 |
Sources
CVE-2026-22773 · published · modified · JSON