Codex Benchmark Details

Codex currently shows benchmark results led by HumanEval (83 / 101, score 28.81), HumanEval pass@10 (15 / 18, score 46.81), HumanEval pass@100 (18 / 21, score 72.31).

Benchmark Results

Codex

Benchmark Results

Coding and Software Engineer

1 evaluations

Benchmark / mode

Score

Rank/total

Standard Mode

28.81

83 / 101

Other

2 evaluations

Benchmark / mode

Score

Rank/total

HumanEval pass@100

Standard Mode

72.31

18 / 21

HumanEval pass@10

Standard Mode

46.81

15 / 18

Compare with other models