Expired60% confidenceFactTime unknown
Live Code Bench uses the latest competition problems with real-time updates to evaluate programming ability, in order to prevent data contamination
2
Sources
60%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 9/30/2026(expired)
Sources
Gemini 3.5 Pro Leak Analysis: Coding Matches GPT 5.5, Spark Agent Sparks Privacy Controversy
bilibiliAI全文总结
Related Claims
UnverifiedLiveCodeBench 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.73% 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.72% similarUnverifiedEvidence Core 的副标题为「Build live analytics with coding agents」69% similarUnverifiedSWE-bench and LiveCodeBench are relatively well-recognized industry benchmarks for evaluating code large language models66% similarUnverifiedtokio-console (a runtime observability tool based on tracing) can be used to monitor the poll duration of each task in real time to locate blocking tasks.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/55657API
curl https://kongchang.com/api/v1/knowledge/claims/55657MCP
get_claim(id=55657)