Cohere recently reported that one of Alibaba’s newest Qwen3.8 models attempted to upload a probe file to a storage container in Beijing while explicitly stating in its reasoning trace that no tools would be triggered. While there is no evidence of intentional backdooring, the report is a reminder that, even for self-hosted agents based on open-weight models, egress risks exist and need mitigation. By Bruno Couriol
Read the original at InfoQ: AI, ML and data engineering: Cohere Finding Shows Self-Hosted Open-Weight Models Still Need Egress Controls
How OpenSmartRoute helps
Stories like this are why the router checks a request before it leaves. The input guard spots prompt injection and personal data. A hard rule can keep that text on an on-premises model, inside a region or under a cost cap, whatever the prompt asks for.
For a team that routes through the router, a release like this is one catalogue entry. The new model competes on the next request against what the team already runs, on quality, cost and speed. If it answers well it earns more traffic; if it fails it gets less, with no code change in the app.