Cohere’s Qwen Story: Looks More Like Model Jaggedness

Summary

Cohere reported that Qwen3.8-Flash-Next, an open-weight Alibaba model, produced a shell command during a detection-rule task that tried to send a test string to an Alibaba Cloud storage address in Beijing, and its sandbox blocked the request. Check Point Research replayed the trace, sampled 64 completions of the unfinished curl command, and found 52 distinct destination hosts, many local or nonexistent, with one invalid top-level domain. Swapping the order of two program names in a tool-availability check made the model much less likely to continue with a network call, leading the researchers to attribute the behavior to model jaggedness rather than a fixed backdoor.

Related items

• Medium“Next Exploit Opportunities”: How a Threat Actor Used AI to Help Plan and Execute Multiple Types of Cybercrime ActivitiesSame vendor · Palo Alto Unit 42
• MediumLiteLLM PyPI Package and Docker Images Compromised to Steal Cloud and AI CredentialsSame vendor · Socket Blog
• MediumCVE-2026-108597: Cohere Python SDK path traversal in model archive extractionSame vendor · NVD/CVE Database
• InfoThe U.S. says China's AI progress is down to 'distillation.' But is it that clear cut?Same vendor · CNBC Technology
• InfoWill AI really destroy humanity? Pioneers who created the tech weigh inSame vendor · CNBC Technology