Scaling the Horizon, Not the Parameters: Reaching Trillion-Parameter Performance with a 35B Agent

Lei Bai, Zongsheng Cao, Yang Chen, Zhiyao Cui, Shangheng Du et al.|June 29, 2026arXiv

Key Takeaway

You can match trillion-parameter model performance on complex reasoning tasks with a 35B model by focusing on longer, more complex action sequences and multi-domain expertise rather than raw parameter count.

Summary

Agents-A1 is a 35B parameter model that achieves trillion-parameter performance on complex tasks by scaling agent horizon—the length and complexity of action sequences—rather than model size.

agents training scaling

Key Terms

agentic-language-model long-horizon-tasks mixture-of-experts multi-teacher-distillation domain-routed-distillation