MirrorCode: how far can coding agents work on their own?
MirrorCode tests whether a coding agent can reimplement a complete program under strict end-to-end tests and project-scale resource budgets.
AI model and agent benchmark analysis
Sourced analysis of AI model and agent benchmarks: what each evaluation measures, what the results show, and where the evidence stops. The first collection focuses on coding agents.
Explore the coding-agent chartArticles
2 sourced benchmark summaries
SlopCodeBench: coding agents degrade on long tasks
SlopCodeBench follows agents as they repeatedly extend their own code, measuring correctness, cost, structural erosion, and verbosity at each checkpoint.
Method
Sources
Each note starts with the benchmark paper or maintained source page. Material claims link to those primary sources.
Changing results
Leaderboard values are paired with their observation date and named configuration. They can change after publication.
Limits
Methodology limits and interpretation are kept near the results they qualify. Benchmark performance is not treated as a general production claim.