Skip to content

Benchmark

Arena Agent Bash Recovery Steps

Agentic benchmarks measure how well models carry out tasks that take several steps. Higher scores are better.

Models measured
50
Scores recorded
52
Sources
1
IQ weight
2.12

Score distribution

0.008
Mean
0.023
Median
0.091
Std dev
-0.075
10th percentile
0.096
90th percentile
-0.256
Min
-0.2560.15
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.12 for chat models, measured over 50 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

Showing 51–52 of 52 results.

Data credits: LMArena, CC BY 4.0 (licence)