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 5 · SOTA2 Research

Long-context Understanding

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

90.5Score

4

Jun 9, 2026

94.3Score

4

Jun 9, 2026

95Score

4

Jun 9, 2026

LongMemEval M 500K

40.2Score

4

Jun 9, 2026

LongMemEval S 125K

82LongMemEval Score

4

Jun 9, 2026

LongBench 493K v2 (L)

70Score

4

Jun 9, 2026

LongBench M 179K v2

68.9Overall Score

4

Jun 9, 2026

LongBench S 46K v2

70.2Score

4

Jun 9, 2026

LongBench-Pro English

37.25Accuracy

4

May 28, 2026

Certified Quantized Attention

6.9Delta Accuracy

4

May 21, 2026

Runtime-Certified Bounded-Error Quantized Attention

0.015Delta Accuracy

4

May 21, 2026

32.03Single-Doc QA Score

4

May 15, 2026

LongBench 32k 2024

57.62HotpotQA Score

4

May 11, 2026

3.08MAT

4

Apr 30, 2026

HELMET shorter context lengths ≤32K

59.34Score (8K Context)

4

Apr 10, 2026

RULER shorter context lengths ≤32K

86.43Performance Score (8K Context)

4

Apr 10, 2026

Spec prefill with Llama-3.2-1B-Instruct

59.63HotpotQA Accuracy

4

Mar 4, 2026

50.48Accuracy (0–4k Context)

4

Mar 4, 2026

95.94RULER Accuracy (4K)

4

Feb 26, 2026

33.9Accuracy

4

Feb 26, 2026

0.609Score

4

Feb 26, 2026

RULER 8K context length

Qwen 1.5 (1.8B)

66.5Avg Task Score

4

Feb 26, 2026

49.11FullKV Performance

2

Jun 16, 2026

46.53Overall Score

2

Jun 12, 2026

Long Context Evaluation

ConceptMoE-A2.5B-60B

80.7Needle Score

2

Feb 26, 2026

Page 5 of 6

Popular datasets

LongBench RULER InfiniteBench LV-Eval HELMET

Follow for updates

Join SOTA2 Contact