Skip to content

Benchmark

BABILong qa5 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

65.6%
Mean
75%
Median
27.7%
Std dev
26%
10th percentile
96.9%
90th percentile
23%
Min
23%98%
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-tuneqa5, 128k· BABILong v0 data docs96.9%
~ Mamba (130M) fine-tuneqa5, 128k· BABILong v0 data docs98%
~ RMT (137M) fine-tuneqa5, 128k· BABILong v0 data docs86%
ChatGLM3 6Bchatglm3-6b-128k, qa5, 128k· BABILong v0 data docs26%
GPT-4-0125-previewqa5, 128k· BABILong v0 data docs51%
Llama3-ChatQA-1.5-8B + RAGqa5, 128k· BABILong v0 data docs75%
Phi-3 Medium 128K InstructPhi-3-medium-128k-instruct, qa5, 128k· BABILong v0 data docs69%
Phi-3 Mini 128K InstructPhi-3-mini-128k-instruct, qa5, 128k· BABILong v0 data docs23%

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