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 generation research benchmarks · SOTA2 Research

Long-context generation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

PG-19 32K context length

329.61Throughput

32

Jun 25, 2026

PG-19 24K context length

393.9Throughput

32

Jun 25, 2026

PG-19 16K context length

486.5Throughput

32

Jun 25, 2026

PG-19 8K context length

553.79Throughput

32

Jun 25, 2026

23.94QMSum

22

Mar 10, 2026

Qwen3 Context length (60K)

5.89Throughput Speedup (α)

12

Feb 26, 2026

Qwen3 Context length (50K)

6.02Throughput Speedup (α)

12

Feb 26, 2026

Qwen3 Context length 40K

5.37Throughput Speedup (α)

12

Feb 26, 2026

Qwen3 Context length 30K

4.37Throughput Speedup (α)

12

Feb 26, 2026

Qwen3 Context length 20K

3.73Throughput Speedup (α)

12

Feb 26, 2026

Qwen3 Context length 10K

2.76Throughput Speedup (α)

12

Feb 26, 2026

PG-19 60K context length

6.29Throughput Speedup (micro)

6

Feb 26, 2026

PG-19 50K context length

5.79Throughput Speedup (micro)

6

Feb 26, 2026

PG-19 40K context length

5.46Throughput Speedup (micro)

6

Feb 26, 2026

PG-19 30K context length

4.75Throughput Speedup (micro)

6

Feb 26, 2026

PG-19 20K context length

4Throughput Speedup (micro)

6

Feb 26, 2026

PG-19 10K context length

2.92Throughput Speedup (micro-averaged)

6

Feb 26, 2026

97.627Completion Rate

3

Apr 14, 2026

Popular datasets

PG-19 Qwen3 PG Qwen LongBench

Follow for updates

Join SOTA2 Contact