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

Gartner, a leading analyst firm, predicts that open-weight models will handle 90% of enterprise agentic workflow steps, while frontier LLMs will only manage the remaining 10%. The firm's senior director, Mark Petty, and group vice president, Raymond Paquet, state in their report that open-weight models are able to perform routine tasks at the same level as the cheapest closed models, but at a quarter of the cost.

These routine steps include classification, tool calling, passing, validating, and reconciling data. Typical operational workflows involve hundreds of such routine executions, with only a few requiring frontier reasoning. The analysts suggest that open-weight models can handle hundreds of these routine steps for every one that needs frontier reasoning.

While open models may be more verbose and retry more often, resulting in a saving per task closer to five or six times, they still offer a significant cost benefit. Open-weights preserve control over enterprise workflows, as they prevent sensitive operations from being moved outside of the organization's perimeter. By 2030, 90% of enterprise operational agent workflow steps are expected to execute on open-weight models.

Enterprises are expected to adopt domain-specific open-weight models, as they perform well in scenarios with a small number of highly stable steps running millions of times a month. The value in AI products lies in harnesses - tools that run agents and manage agent fleets. A good harness saves money, imposes strong guardrails, and manages agent fleets effectively.

Frontier models will continue to play a role in research and development, while open-weight models become operational expenses after frontier models have done the heavy lifting.