Skip to content

Benchmark

LOFT · Text Retrieval

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

Models measured
2
Scores recorded
2
Sources
1
IQ weight
No weight published

Score distribution

72.9%
Mean
79.5%
Median
6.6%
Std dev
66.3%
10th percentile
79.5%
90th percentile
66.3%
Min
66.3%79.5%
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1Gemma 4 31B ITGemma 4 Technical Report · Table 9 · without thinking · Recall@k · 128K· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.79.5%
2Gemma 4 26B A4BGemma 4 Technical Report · Table 9 · without thinking · Recall@k · 128K· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.66.3%

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.