Cisco Launches Model Foundry for Partner-Developed Enterprise AI
Cisco has introduced Model Foundry, a platform service within Cisco Cloud Control that allows partners to deploy and customize AI models for enterprise infrastructure operations.
The service provides access to Cisco-developed models, including Deep Network 1.4 for troubleshooting, Antares-3B for vulnerability detection, Decision Model 1.0 for agent decision-making and Cisco Time Series Model 1.0 for observability data.
Partners can evaluate models, deploy them through OpenAI-compatible APIs and fine-tune them using customer-specific configurations, logs, alerts and operational data.
Model Foundry manages GPU resources, training pipelines and model evaluations. Administrative controls include tenant-specific API keys, model access restrictions and usage monitoring.
In laboratory testing with World Wide Technology, a customized Cisco Deep Network Model recommended the correct remediation in 11 of 12 scenarios, compared with eight for a frontier model.
For enterprise IT teams, the platform offers an alternative to general-purpose AI models for specialized infrastructure tasks.
Model Foundry enters controlled availability at Cisco Partner Summit, with partners receiving initial access.
Posted by Pure AI Editors on 10/26/2026