LiveCodeBench (Jain et al., UC Berkeley, MIT and Cornell) collects problems from periodic LeetCode, AtCoder and Codeforces contests, annotates them with release dates so that a model can be scored only on problems published after its training cutoff, and evaluates several code scenarios beyond generation (site). In this corpus it is the yardstick for Apple’s simple self-distillation, which moved Qwen3-30B-Instruct from 42.4% to 55.3% pass@1 on the v6 release with no verifier, teacher or reinforcement learning.
Benchmark
LiveCodeBench
A contamination-free benchmark of LLM coding ability that continuously collects new problems from LeetCode, AtCoder and Codeforces contests and evaluates code generation, self-repair, code execution and test-output prediction.
Sources
Attribution and reporting linked by this record; links do not imply independent verification.
- Welcome to April 5, 2026 (theinnermostloop.substack.com)
Development timeline
Newest reporting dates first. Reporting dates may differ from event dates.
Cite or correct this record
Suggest a correction — opens a GitHub draft for you to review and submit.