Skip to content

Benchmark

Internal API instruction following (hard)

Instruction following benchmarks measure how closely models follow instructions. Scores are percentages.

Models measured
7
Scores recorded
7
Sources
1
IQ weight
1.83

Score distribution

46.1%
Mean
49.1%
Median
11.4%
Std dev
31.6%
10th percentile
54%
90th percentile
29.2%
Min
29.2%64%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.83 for chat models, measured over 7 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1GPT 5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim64%
2GPT-4.5 Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim54%
3o3 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim50%
4GPT 4.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim49.1%
5GPT-4.1 mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim45.1%
6GPT-4.1 nano· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim31.6%
7GPT 4o· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim29.2%

Data credits: Data by LLM Stats