Verified90% confidenceFactExact time
HumanEval was released by OpenAI and contains 164 hand-crafted Python programming problems, with pass@k as its core metric.
10
Sources
90%
Confidence
Long-term
Relevance
7/12/2026
First Seen
Sources
Multi-Model Tiered Deployment: Is an AI Gateway Worth Adopting?
redditr/mlops7/11/2026
Related Claims
UnverifiedHumanEval由164道Python编程题组成75% similarUnverifiedA developer submitted a fully offline-running Python invoice/receipt classification tool proposal in the Hacktoberfest open source AI challenge.63% similarUnverifiedThe HumanEval test set contains only 164 Python programming problems and tests only function-level code completion.62% similarUnverifiedThe author provides a standard-library-only Python script, ask.py, that calls OpenAI-compatible endpoints, appending one line of JSON log per attempt, with key fields including a pinned model identifier, base URL, timestamp, prompt hash, character count, whether usage was reported, input and output token counts, latency in milliseconds, and error information.61% similarUnverifiedThe Humanizer project is primarily written in Python60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/492464API
curl https://kongchang.com/api/v1/knowledge/claims/492464MCP
get_claim(id=492464)