Skip to content

Benchmark

Vals Vibe Code Bench v1.1

Coding benchmarks compare models on programming tasks. Scores are percentages.

Models measured
119
Scores recorded
119
Sources
1
IQ weight
1.95

Score distribution

52.2%
Mean
64%
Median
31.7%
Std dev
5.1%
10th percentile
87.8%
90th percentile
0%
Min
0%93.8%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.95 for chat models, measured over 119 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

Showing 51–100 of 119 results.