Unverified60% confidenceFactExact time
LiveCodeBench continuously crawls new problems from competitive programming platforms such as LeetCode, Codeforces, and AtCoder to prevent models from gaming the benchmark, and its evaluation covers four dimensions: code generation, self-repair, code execution reasoning, and test output prediction.
2
Sources
60%
Confidence
Long-term
Relevance
9/6/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedLiveCodeBench Hard-level problems typically require multi-step reasoning, complex algorithm design, and handling of edge cases, placing extremely high demands on the model's one-shot code correctness rate.76% similarVerifiedLiveCodeBench introduces a timestamp filtering strategy that selects problems published after a model's training cutoff date to ensure that the evaluation data is entirely new to the model being tested.75% similarUnverifiedSWE-bench and LiveCodeBench are relatively well-recognized industry benchmarks for evaluating code large language models74% similarVerifiedMainstream benchmarks for evaluating programming large models include HumanEval, MBPP, SWE-bench, and LiveCodeBench71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/863532API
curl https://kongchang.com/api/v1/knowledge/claims/863532MCP
get_claim(id=863532)