Skip to content

Benchmark

hal-assistantbench

Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages.

Models measured
10
Scores recorded
14
Sources
1
IQ weight
0.91

Score distribution

16.2%
Mean
13.8%
Median
11.1%
Std dev
2.6%
10th percentile
35.2%
90th percentile
2%
Min
2%38.8%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.91 for chat models, measured over 10 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

Tokens the model generated per task, as measured by the source. Comparable within this benchmark only.

Data credits: HAL (Princeton) — cite the HAL paper