LetTokenBurn
Sign in
Auto57%Harbor: pendingCodingAgentic / Tool Use

CORE-Bench

CORE-Bench (EMNLP 2026) Evaluation Code: https://github.

PaperRepoDataset

How to run

ltb · core-bench-hf-bench
$ npm i -g lettokenburn-cli
$ ltb bench info core-bench-hf-bench

External harness. No Harbor adapter has been generated for this benchmark yet. `ltb bench info` prints the known links.

Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs

Scores

0 sourced entries

No sourced scores recorded yet.

Every number links to where it was published. Low-confidence rows are muted and need per-cell verification.

Reproduced by the community

All runs
Loading community runs…

About

CORE-Bench (EMNLP 2026)

Evaluation Code: https://github.com/zhangfw123/CORE-Bench-Eval CORE-Bench is a benchmark for code retrieval in agentic coding settings. It frames retrieval as requirement-driven repository search: given an issue, feature request, refactoring request, or other development query, a retriever must find the code and repository context needed by a coding agent. The paper defines three evaluation levels:

Level-1: Code Understanding evaluates traditional…