CVE-2026-54233Medium· 6.5▾ SunlitvLLM: OOM Denial of Service via Audio Decompression Bomb
▾ Sunlit zone — Low / medium · no exploitation signal
impact 35.8 · likelihood 0.1 · exploitation 0
Need a working PoC? Pro members can cast a request and our team develops one — it lands right here.
Exploit-prediction probability, daily snapshots since Jul 31.
Disclosure to exploitation, from the record and what we observed since indexing it.
Disclosed via OSV
0.4%
Last analysed / modified upstream
vLLM's /v1/audio/transcriptions endpoint limits compressed upload size but not decoded PCM output. A 25MB OPUS file expands to ~14.9GB of float32 PCM at decode time. Tested on vLLM v0.19.0.
SpeechToTextProcessor rejects uploads over VLLM_MAX_AUDIO_CLIP_FILESIZE_MB (default 25MB) based on compressed byte length, but the audio decoder in audio.py accumulates all decoded frames into memory with no size limit before returning:
# speech_to_text.py L184-189
if len(audio_data) / 1024 ** 2 > self.max_audio_filesize_mb:
raise VLLMValidationError(...)
y, sr = load_audio(buf, sr=self.asr_config.sample_rate) # decoded size unchecked
# audio.py L77-107
chunks: list[npt.NDArray] = []
for frame in container.decode(stream):
chunks.append(frame.to_ndarray())
audio = np.concatenate(chunks, axis=-1).astype(np.float32) # single contiguous allocation
A 25MB OPUS file at 6kbps encodes ~8.7 hours of audio. Decoding produces ~5.7GB of float32 PCM (232x amplification), and np.concatenate then allocates a second contiguous array, bringing peak RSS to ~14.9GB from a single request. SpeechToTextConfig.max_audio_clip_s (default 30s) applies only after the full decode and does not prevent the allocation.
An unauthenticated attacker can exhaust server memory with a small number of concurrent requests, each a valid upload within the documented size limit. Severity was assessed with reference to prior OOM vulnerability reports in vLLM.
A fix for this vulnerability was merged here: https://github.com/vllm-project/vllm/pull/44970
vllm < 0.24.0Upgrade to a patched release:
vllm 0.24.0Connected by shared product, vendor, weakness, or advisory.
CVE-2026-105760Medium· 5.3vLLM is an inference and serving engine for large language models
CVE-2026-105758Medium· 5.3vLLM is an inference and serving engine for large language models
CVE-2026-105754Medium· 6.5vLLM is an inference and serving engine for large language models
CVE-2026-105756Medium· 6.5vLLM is an inference and serving engine for large language models
CVE-2026-105752Low· 3.1vLLM is an inference and serving engine for large language models
CVE-2026-69147Medium· 6.5vLLM is an inference and serving engine for large language models