Open models can do 90 percent of agentic work at a twentieth of the cost of a frontier LLM: Gartner

A Gartner analysis by Mark Petty and Raymond Paquet argues that enterprise agentic AI workflows involve mostly routine steps—classification, tool calling, data validation and reconciliation—which open-weight models handle at cost parity with frontier models but for significantly less money. The firm estimates open models cost a quarter as much as the cheapest closed alternative and as little as a fifth the cost of frontier models on routine tasks. Even accounting for open models being more verbose and requiring more retries, the savings drop to five or six times rather than twenty times, still representing substantial cost advantages for routine operational tasks.

Summary and key points written by AI from the headline and a short excerpt, so they can be incomplete or wrong. The Register did the reporting; read the original for the full story.