Kwai-Klear

GoLongRL-4B

Deploy Dedicated

📌 Overview

Overall performance comparison on long-context benchmarks (DocMath, LongBench-V2, Frames, MRCR, CorpusQA, LBV1-QA).

GoLongRL-4B achieves strong long-context performance at the 4B scale. Our framework combines the following:

Capability-Oriented Dataset (23K samples, 9 task types). Guided by a taxonomy of long-context capabilities, the dataset covers precise retrieval, comprehension, exhaustive retrieval, numerical reasoning, structured extraction, structured matching, graded ranking, sequence ordering, and summarization. Each task is paired with its natural evaluation metric (EM, Accuracy, F1, math_verify, IoU, SubEM, NDCG, Pairwise, ROUGE-L) as the reward function.
TMN-Reweight. To address optimization challenges from heterogeneous rewards, we propose TMN-Reweight, which combines task-level mean normalization for cross-task reward scale alignment with difficulty-adaptive weighting for more reliable advantage estimation. It provides a modest but consistent improvement over vanilla GRPO.
Full Open Release. We publicly release the complete dataset, the four-phase construction pipeline, and all training code.

Key Results

Under the same vanilla GRPO setup, our dataset alone outperforms the closed-source QwenLong-L1.5 dataset.
TMN-Reweight further improves average performance to 63.0 at 4B scale, surpassing QwenLong-L1.5 with its specialized AEPO algorithm.
General capabilities (MMLU-Pro, AIME24/25, GPQA) are preserved or improved, with substantial gains on dialogue memory (LongMemEval +13.6) and agentic memory benchmarks.

🔍 Evaluation

Evaluation uses QwenLong-Benchmarks, covering three capability dimensions:

Table with columns: Dimension, Benchmarks
Dimension	Benchmarks
Long-Context	LongBench-V2, MRCR (≤128K / 128K–512K / 512K–1M), Frames, LongBench QA, DocMath, CorpusQA (≤128K / ≤1M)
General	MMLU-Pro, AIME 2024/2025, GPQA-Diamond
Memory	BFCL-V4 (memory subset), LongMemEval

🤝 Citation

bibtex

@misc{lv2026golongrlcapabilityorientedlongcontext,      title={GoLongRL: Capability-Oriented Long Context Reinforcement Learning with Multitask Alignment},       author={Minxuan Lv and Tiehua Mei and Tanlong Du and Junmin Chen and Zhenpeng Su and Ziyang Chen and Ziqi Wang and Zhennan Wu and Ruotong Pan and jian Liang and Ruiming Tang and Han Li},      year={2026},      eprint={2605.19577},      archivePrefix={arXiv},      primaryClass={cs.CL},      url={https://arxiv.org/abs/2605.19577}, }

Available on FriendliAI

Dedicated Endpoints

Run this model inference on single tenant GPU with unmatched speed and reliability at scale.

Learn more

Model Details

Model Provider

Kwai-Klear

Model Tree

Base

this model

Input Modalities

Text

Output Modalities

Text

Supported Functionality

Dedicated Endpoints

Explore FriendliAI today

Get started Talk to an engineer

📌 Overview

Overall performance comparison on long-context benchmarks (DocMath, LongBench-V2, Frames, MRCR, CorpusQA, LBV1-QA).

GoLongRL-4B achieves strong long-context performance at the 4B scale. Our framework combines the following:

Capability-Oriented Dataset (23K samples, 9 task types). Guided by a taxonomy of long-context capabilities, the dataset covers precise retrieval, comprehension, exhaustive retrieval, numerical reasoning, structured extraction, structured matching, graded ranking, sequence ordering, and summarization. Each task is paired with its natural evaluation metric (EM, Accuracy, F1, math_verify, IoU, SubEM, NDCG, Pairwise, ROUGE-L) as the reward function.
TMN-Reweight. To address optimization challenges from heterogeneous rewards, we propose TMN-Reweight, which combines task-level mean normalization for cross-task reward scale alignment with difficulty-adaptive weighting for more reliable advantage estimation. It provides a modest but consistent improvement over vanilla GRPO.
Full Open Release. We publicly release the complete dataset, the four-phase construction pipeline, and all training code.

Key Results

Under the same vanilla GRPO setup, our dataset alone outperforms the closed-source QwenLong-L1.5 dataset.
TMN-Reweight further improves average performance to 63.0 at 4B scale, surpassing QwenLong-L1.5 with its specialized AEPO algorithm.
General capabilities (MMLU-Pro, AIME24/25, GPQA) are preserved or improved, with substantial gains on dialogue memory (LongMemEval +13.6) and agentic memory benchmarks.

🔍 Evaluation

Evaluation uses QwenLong-Benchmarks, covering three capability dimensions:

Table with columns: Dimension, Benchmarks
Dimension	Benchmarks
Long-Context	LongBench-V2, MRCR (≤128K / 128K–512K / 512K–1M), Frames, LongBench QA, DocMath, CorpusQA (≤128K / ≤1M)
General	MMLU-Pro, AIME 2024/2025, GPQA-Diamond
Memory	BFCL-V4 (memory subset), LongMemEval

🤝 Citation

bibtex

@misc{lv2026golongrlcapabilityorientedlongcontext,      title={GoLongRL: Capability-Oriented Long Context Reinforcement Learning with Multitask Alignment},       author={Minxuan Lv and Tiehua Mei and Tanlong Du and Junmin Chen and Zhenpeng Su and Ziyang Chen and Ziqi Wang and Zhennan Wu and Ruotong Pan and jian Liang and Ruiming Tang and Han Li},      year={2026},      eprint={2605.19577},      archivePrefix={arXiv},      primaryClass={cs.CL},      url={https://arxiv.org/abs/2605.19577}, }

GoLongRL-4B

README

📌 Overview

Key Results

🔍 Evaluation

🤝 Citation

Explore FriendliAI today

README

📌 Overview

Key Results

🔍 Evaluation

🤝 Citation