Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Long-context Understanding research benchmarks – Page 3 · SOTA2 Research

Long-context Understanding

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

RULER 8K context

Training–Inference Consistent Segmented Execution

54.15CWE

13

May 13, 2026

96.1Accuracy (4K Context)

13

Feb 26, 2026

LongBench, LongBench v2 & LooGLE Combined

45.71Overall Score

12

Jul 8, 2026

57CR Score

12

Jul 8, 2026

1.61Decoding Speedup

12

Jun 30, 2026

1.48Decoding Speedup

12

Jun 30, 2026

1.64Decoding Speedup

12

Jun 30, 2026

Long-document Understanding and Multi-turn Memory Overall

83.8Accuracy

12

Jun 30, 2026

LongBench (eval)

Fixed Sparse (16, 0.02)

39.6Qasper Score

12

May 11, 2026

215P95 Latency (ms)

12

Apr 23, 2026

27.77NQA Score

12

Apr 10, 2026

57.33Accuracy

12

Apr 21, 2026

LongBench 64K context length

63.8Average Score

11

Jul 8, 2026

Open Long Context Benchmarks

55.77Loogle Score

11

Apr 10, 2026

LongBench TriviaQA

86.67F1 Score

10

Jun 11, 2026

LongBench QASPER

45.14F1 Score

10

Jun 11, 2026

LongBench NarrativeQA

27.4F1 Score

10

Jun 11, 2026

LongBench MuSiQue

22.32F1 Score

10

Jun 11, 2026

LongBench MFQA-zh

14.79F1 Score

10

Jun 11, 2026

LongBench MFQA-en

47.18F1 Score

10

Jun 11, 2026

LongBench HotpotQA

52.15F1 Score

10

Jun 11, 2026

LongBench 2WikiMQA

37.47F1 Score

10

Jun 11, 2026

LongBench Short v2

61.66Average Score @4 (Short Context)

10

May 27, 2026

52.28Single Doc QA

10

Feb 26, 2026

RULER 64K context 1.0 (test)

100N-S Score 1

9

May 18, 2026

Page 3 of 6

Popular datasets

LongBench RULER InfiniteBench LV-Eval HELMET

Follow for updates

Join SOTA2 Contact