Skip to content

Benchmark

BABILong qa2 16k

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

Models and systems measured
13
Scores recorded
14
Sources
1
IQ weight
No weight published

Score distribution

42.5%
Mean
32%
Median
31.3%
Std dev
13%
10th percentile
98%
90th percentile
9%
Min
9%100%
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-tuneqa2, 16k· BABILong v0 data docs100%
~ Mamba (130M) fine-tuneqa2, 16k· BABILong v0 data docs98%
~ RMT (137M) fine-tuneqa2, 16k· BABILong v0 data docs94%
ChatGLM3 6Bchatglm3-6b-128k, qa2, 16k, 128k· BABILong v0 data docs31%
GPT-4-0125-previewqa2, 16k· BABILong v0 data docs59%
Llama 2 7bLLaMA-2-7B-32K, qa2, 16k, 32k· BABILong v0 data docs21%
Llama3-ChatQA-1.5-8B + RAGqa2, 16k· BABILong v0 data docs13%
Mistral 7B Instruct v0.2Mistral-7b-Instruct-v0.2, qa2, 16k· BABILong v0 data docs9%
Mixtral 8x22B Instruct v0.1Mixtral-8x22B-Instruct-v0.1, qa2, 16k· BABILong v0 data docs41%
Mixtral 8x7B Instruct v0.1Mixtral-8x7B-Instruct-v0.1, qa2, 16k· BABILong v0 data docs18%
Phi-3 Medium 128K InstructPhi-3-medium-128k-instruct, qa2, 16k· BABILong v0 data docs41%
Phi-3 Mini 128K InstructPhi-3-mini-128k-instruct, qa2, 16k· BABILong v0 data docs23%
Yi-34B01-ai/Yi-34B-200k, qa2, 16k, 200k· BABILong v0 data docs32%
Yi-34B01-ai/Yi-34B, qa2, 16k· BABILong v0 data docs15%

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