PenguinScrolls
Hunyuan-A13B-Instruct 87.7% 
Qwen2.5-72B 81.8% leads Hunyuan-A13B-Instruct 87.7% leads since June 2025 Models answer user-style questions over long natural documents such as books, financial reports and legal papers, up to 128K tokens. Tasks cover extraction, localization, qualitative analysis and numerical reasoning.
The release has 1,316 single-turn items in four categories. Tencent reports it in the Hunyuan-A13B technical report.
- Made by
- Tencent Hunyuan
- Released
- November 2024
- Kind
- Reasoning over long documents
- Level
- Everyday
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- All public
- Strongest model tested
- Qwen2.5-72B #165 on the capability index
- Last in a system card
- June 2025
- Last model added
- June 2025
- Scores refreshed
- November 12th, 2026
Hunyuan-A13B-Instruct leads with 87.7%
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
Qwen2.5-72B leads with 81.8% across all independent runs
Best score per model across the official leaderboard, Epoch AI’s own runs and any board Epoch AI copied. Each bar and row names its source.
Hunyuan-A13B-Instruct leads with 87.7%, claims included
Best score per model from every source, including labs’ own numbers and claims Epoch AI collected from reports and papers. Labs pick their own settings, so a claim can beat an independent run of the same model.
Source: Labs’ model cards
- 1 Qwen2.5-72B 81.8%
- 2
Llama 3.1-70B 66.5%
Source: Official PenguinScrolls leaderboard
- 1 Hunyuan-A13B-Instruct 87.7%
- 2 Qwen2.5-72B 81.8%
- 3 Llama 3.1-70B 66.5%
Sources: Labs’ model cards and official PenguinScrolls leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Hunyuan-A13B-Instruct | 87.7% | Jun 2025 | Lab model card |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Qwen2.5-72B | 81.8% | Sep 2024 | Official leaderboard | |
| 2 | Llama 3.1-70B | 66.5% | Jul 2024 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Hunyuan-A13B-Instruct | 87.7% | Jun 2025 | Lab model card | |
| 2 | Qwen2.5-72B | 81.8% | Sep 2024 | Official leaderboard | |
| 3 | Llama 3.1-70B | 66.5% | Jul 2024 | Official leaderboard |
The benchmark's own leaderboard has 5 entries, 2 of them models we track. Switch the scores above to rank them together.
Tencent’s own number for Hunyuan-A13B-Instruct passed 85% in June 2025
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Sources: Official PenguinScrolls leaderboard and labs’ model cards
Labs report it only in system cards
We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark. Some system cards do, below.
One system card reports it
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| Jun 26, 2025 | Hunyuan-A13B-Instruct technical report | 87.7 | PenguinScrolls (Table 5) | Capabilities |
More in Long context
- MRCR v2 In use Multi-round coreference: retrieve the n-th of several similar requests hidden in a long synthetic conversation, up to 1M tokens. Muse Spark 1.3 98.5%
- LongBench v2 In use 503 multiple-choice questions over 8K-2M-word contexts (QA, dialogue, code repos, structured data). Accuracy.
Inkling 2 Small 70.6% - GraphWalks In use A graph is given as an edge list in long context. The model answers BFS or parent queries, scored by F1. Claude Fable 5 99.96%
- OpenAI MRCR v2 In use Models read a long synthetic chat that hides 2, 4 or 8 identical writing requests and must reproduce the response to the i-th one. Scored by mean string match ratio per context bin.
Gemini 3.7 Flash 89% - CL-bench In use Model must learn new knowledge from a supplied long context (rules, procedures, data) and solve tasks. Strict LLM-verified rubrics. GPT-5.4 27.9%
- AA-LCR 1.1 In use Models answer 100 questions that need reasoning across sets of real documents of about 100K tokens. An LLM equality checker with a fixed system prompt compares each answer with the key.
Step 5 Preview 88.3%