Skip to content

Benchmark

BABILong qa3 128k

Long context benchmarks measure how well models use information in long inputs. Scores are percentages. Higher is better.

Models and systems measured
8
Scores recorded
8
Sources
1
IQ weight
No weight published

Score distribution

35.5%
Mean
32%
Median
29.5%
Std dev
9%
10th percentile
81%
90th percentile
4%
Min
4%86%
The band covers the middle 80% of scores. The line marks the median.

Results

The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.

Model or systemScore
~ ARMT (137M) fine-tuneqa3, 128k· BABILong v0 data docs86%
~ Mamba (130M) fine-tuneqa3, 128k· BABILong v0 data docs81%
~ RMT (137M) fine-tuneqa3, 128k· BABILong v0 data docs36%
ChatGLM3 6Bchatglm3-6b-128k, qa3, 128k· BABILong v0 data docs9%
GPT-4-0125-previewqa3, 128k· BABILong v0 data docs32%
Llama3-ChatQA-1.5-8B + RAGqa3, 128k· BABILong v0 data docs19%
Phi-3 Medium 128K InstructPhi-3-medium-128k-instruct, qa3, 128k· BABILong v0 data docs17%
Phi-3 Mini 128K InstructPhi-3-mini-128k-instruct, qa3, 128k· BABILong v0 data docs4%

Data credits: RMT-team, Kuratov et al. Official v0 numeric CSV; results licence not separately stated.