Cisco’s Tiny AI Models Just Beat GPT-5.5 on Cost by 172x
Today’s Techstrong Gang panel — Alan Shimel, Kate Scarcella, Sid Nag, and Chris Blask — unpacks three major stories, including how Cisco AI models are changing enterprise security economics. First, there’s the rise of open-weight AI models. Then, Cisco’s ultra-cheap security AI models take center stage. Finally, Anaconda pushes further into agentic AI development.
Free as in Freedom, Free as in Beer: How the AI Model War Ends
Alan Shimel argues that AI providers risk becoming gatekeepers. Additionally, AI intelligence itself is becoming a commodity. Capability keeps spreading into cheaper, smaller open-weight AI models. Meanwhile, Microsoft is positioning to control the layer above the models: Azure, Foundry, and Copilot. As a result, it can profit regardless of which lab has the best model in a given month. Shimel frames “free as in beer” (commodity-priced intelligence) and “free as in freedom” (user control over where models run and what data they retain) as two forces. Together, these forces undercut proprietary AI’s pricing power. Read the full piece on Techstrong.ai.
Cisco Antares: Cheap, Open-Weight Security AI Models
Cisco unveiled Cisco Antares on July 21. These are small, open-weight AI models built to hunt for known vulnerabilities in software codebases. Antares-350M and Antares-1B are now available on Hugging Face to verified cyber defenders. However, a larger Antares-3B is held back for Cisco’s own products. According to Cisco VP and chief AI scientist Amin Karbasi, Antares scanned 500 repositories in about 15 minutes for under $1. In contrast, the same task took roughly five hours and $100+ using GPT-5.5. Therefore, Cisco claims a 172x cost advantage. Details via Security Boulevard.
Anaconda Acquires Kilo Code
Anaconda acquired Kilo Code, an open-source, model-agnostic AI coding agent. Furthermore, more than 3 million developers already use it. Kilo remains MIT-licensed and gives access to 500+ AI models with no markup or lock-in. Overall, the deal signals Anaconda’s push to become an AI-native development platform beyond its Python roots. More at DevOps.com.
Watch daily on Techstrong TV for more DevOps, AI, and enterprise tech coverage.
