Open-Weight Models Set to Dominate 90 Percent of Enterprise Agentic Workflows
Cost efficiency and data control drive the shift toward open-weight models for routine tasks.
1Routine Tasks Shift to Open-Weight
Gartner analysts Mark Petty and Raymond Paquet report that 90 percent of agentic workflow steps are routine. These tasks, including data classification, tool calling, and validation, can be handled by open-weight models. By shifting these repetitive executions away from frontier models, enterprises can optimize their operational efficiency and reduce infrastructure overhead significantly.
2Cost Advantage Over Frontier Models
Open-weight models currently perform routine tasks at a quarter of the price of the cheapest closed models. When compared to frontier models, the cost drops to as little as one-twentieth. While verbosity and retry rates may adjust real-world savings to five or six times, the financial incentive remains a compelling driver for adoption.
3Data Control and Security Benefits
Beyond cost, open-weight models offer superior data sovereignty. Sending routine steps to frontier model APIs risks moving proprietary workflow history outside the enterprise perimeter. By keeping these processes in-house, companies maintain control over their operational data, which is a critical component of the value proposition for modern SaaS-based business architectures.
4Strategic Planning for 2030
The analysts project that by 2030, 90 percent of enterprise operational agent workflow steps will execute on open-weight models. This transition is expected to favor domain-specific models that excel in scenarios where a small number of highly stable steps are executed millions of times each month, effectively commoditizing the underlying model layer.
5The Role of AI Harnesses
The true value in AI products is shifting toward harnesses, the tools that manage and route agent fleets. A sophisticated harness intelligently directs work to the most appropriate model without user intervention. These tools are essential for imposing guardrails, managing costs, and ensuring that the right model handles the right task.
6Frontier Models as Capital Expense
Frontier models will retain a vital role as the engine for research and development within organizations. They function effectively as a capital expense, providing the heavy lifting required for complex reasoning. Once these capabilities are established, the operational execution of routine tasks is then offloaded to more cost-effective open-weight alternatives.
7Operationalizing Agentic Workflows
Enterprises are moving toward a hybrid model where frontier intelligence and open-weight efficiency coexist. This architecture allows firms to balance high-level reasoning with high-volume, low-cost execution. By treating models as commodities, businesses can focus on building robust agentic systems that scale without incurring the prohibitive costs associated with constant frontier model usage.
8Implications for Enterprise Strategy
The shift toward open-weight models represents a fundamental change in how companies approach AI deployment. By prioritizing control and cost-efficiency, organizations can protect their intellectual property while scaling automation. The long-term success of these initiatives will depend on the quality of the orchestration layer and the ability to route tasks effectively.