TrueFoundry Launches TrueForge, Claiming Up to 75% Lower AI Agent Costs
Updated
Updated · InfoWorld · Aug 20
TrueFoundry Launches TrueForge, Claiming Up to 75% Lower AI Agent Costs
3 articles · Updated · InfoWorld · Aug 20
Summary
TrueFoundry unveiled TrueForge, an open-source AI agent harness that runs on enterprise infrastructure and supports OpenAI, Anthropic and 20-plus other models.
In a 14-task DevRev Enterprise-Bench test, TrueFoundry said TrueForge matched Claude Managed Agents at about 11 completed tasks but cut per-run cost to $8.50 from $11.80 with Opus 4.8, and to $2.90 with GLM-5.2.
Those results imply roughly 30% to 75% savings, but the benchmark was run by TrueFoundry and has not been independently validated in broader production workloads.
Analysts said separating the agent runtime from the model provider can reduce vendor lock-in and help regulated companies add their own controls, though self-hosting shifts maintenance, compliance and monitoring burdens onto enterprises.
Whether the model works financially will depend on workload volume, token use and in-house infrastructure, with experts saying self-hosting is more attractive for high-volume agents than for smaller or unpredictable usage.