Study
Agentic AI · Agent evaluation
Evaluating agents means checking the final answer and the trajectory: task success, tool-call correctness, wasted steps and cost, scored here over a few recorded agent runs.
- Agents can reach a right answer by a wrong route, or a wrong answer by a sensible one, so both the outcome and the trajectory need checking.
- Benchmarks like GAIA and SWE-bench score task success; production evals also track steps, cost and failures.