Bedrock AgentCore Evaluations scores agents from OpenTelemetry spans, so the framework stops mattering.

It reconstructs sessions from three OTel span types, invoke agent, inference and execute tool, then applies the same scoring regardless of SDK (AWS). Strands, LangGraph, OpenAI Agents SDK, LlamaIndex, Google ADK and the Claude Agent SDK are named explicitly, with generic classification for anything emitting spans under opentelemetry.instrumentation.* or openinference.instrumentation.*. Built-in evaluators are GoalSuccessRate at session level and Correctness and Helpfulness at trace level. No pricing or regions published yet.