用任务结果、决策等价、约束保留和单个成功任务成本,验收 Agent 上下文压缩,而不是只看 Token 节省率。
Evaluate AI agent context compression with task outcomes, decision equivalence, constraint retention, and cost per successful task.
Astra scored 62.7% or 99.9% on ARC-AGI-3 under different harnesses. Here is what changed, what the numbers mean, and what remains unproven.