MirrorCode 基准测试：Claude Opus 4.7 以 56% 解决率领先，但最复杂任务仍失败
An AI model programmed nonstop for 19 days on a single MirrorCode task that cost $2,600 to run
Epoch AI 搞了个新基准 MirrorCode，专测 AI 能不能凭空抄作业。Claude Opus 4.7 解了一半，但最难的题全挂，甚至有个模型烧了 19 天才花掉 2600 刀。
Epoch AI 发布新基准 MirrorCode，测试 AI 模型能否在无原始代码时重建完整程序。Claude Opus 4.
7 以 56% 的解决率领先，曾在 14 小时内重建 16,000 行工具包。个别模型为单个 MirrorCode 任务连续运行 19 天，花费 2,600 美元。
所有测试模型在最复杂任务上均失败。
An AI model programmed nonstop for 19 days on a single MirrorCode task that cost $2,600 to run
Epoch AI's new MirrorCode benchmark tests whether AI models can recreate complete programs without access to the original code. Claude Opus 4.7 leads with a 56 percent solve rate, rebuilding a 16,000-line toolkit in just 14 hours. But every model tested still fails on the most complex tasks. The article An AI model programmed nonstop for 19 days on a single MirrorCode task that cost $2,600 to run appeared first on The Decoder .
This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include MirrorCode, Epoch AI, Claude Opus 4.7, 代码生成.
Source attribution: curated from Decoder (@Matthias Bastian). Cross-source corroboration: 2 independent sources have covered this development (including Epoch AI), confirming industry-wide signal strength.
- Epoch AI06-27 04:56Source
