# Meta 发布 WildArtifactBench 评测框架

- 来源：AI at Meta (@AIatMeta)
- 发布时间：2026-08-21 02:25
- AIHOT 分数：53
- AIHOT 链接：https://aihot.virxact.com/items/cmt1v1x4p09gsroovhe8zt77y
- 原文链接：https://x.com/AIatMeta/status/2090505413817246050

## AI 摘要

Meta 发布内部评测框架 WildArtifactBench，用于评估多模态智能体在复杂真实任务中的表现。该框架采用人类与智能体偏好评判的胜率和 Elo 评分，替代严格的标准答案评分规则，以覆盖更多实际多模态工作流。Meta 已开源其中 10 个任务。

## 正文

Today we’re also previewing WildArtifactBench, an internal evaluation framework designed to assess agents on complex, real-world tasks across diverse deliverable formats.

By using win rates and Elo scores from human and agentic preference judges rather than strict ground-truth rubrics, it expands task coverage across practical multimodal workflows.

We’re releasing 10 tasks from WildArtifactBench as a step forward in our ability to measure the real practical utility delivered by multimodal agents: https://research.meta.ai/wild-artifact-bench