PeerQA
Models answer peer reviewers' questions about a scientific article using evidence from the article.
The original authors supplied reference answers to the reviewers' questions.
- Made by
- UKP Lab
- Released
- February 2025
- Kind
- Research science
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Items from
- Written by experts, Real work and use
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Biology and medicine
- Strongest model tested
-
Gemma 3 27B #165 on the capability index
- Last model added
- May 2025
- Scores refreshed
- November 20th, 2026
Mistral-7B-Instruct-v0.2 leads with 0.24 among models released by May 2025 Mistral-7B-Instruct-v0.2 leads with 0.24 among models released by May 2025
Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. The headline is answer-generation ROUGE-L against the authors’ answer evidence, the first measure in Table 12. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.
The newest model with a score here came out in May 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 0.24
Mistral-7B-Instruct-v0.2
- 0.20 GPT-4o
- 0.18 Llama-3 IT-8B-32k
- 0.18
Llama 3.3 70B Instruct
- 0.18 Llama-3 IT-8B-8k
- 0.17 deepseek-r1-0528-qwen3-8b
- 0.17 DeepSeek-R1-Llama-8B-128k
- 0.17
DeepSeek-R1-Distill-Qwen-32B
- 0.16
DeepSeek-R1-Distill-Qwen-14B
- 0.16
DeepSeek-R1-Distill-Qwen-7B
- 0.16
Gemma 3 27B
- 0.15 GPT-3.5 Turbo
- 0.15
DeepSeek-R1-Distill-Llama-70B
- 0.13 Command-R v01-34B-128k
- 1
Mistral-7B-Instruct-v0.2 Gold evidence 0.24
- 2 GPT-4o Aug 2024, Gold evidence 0.20
- 3 Llama-3 IT-8B-32k Full text 0.18
- 4
Llama 3.3 70B Instruct Full text 0.18
- 5 Llama-3 IT-8B-8k Top 20 paragraphs 0.18
- 6 deepseek-r1-0528-qwen3-8b Gold evidence 0.17
- 7 DeepSeek-R1-Llama-8B-128k Gold evidence 0.17
- 8
DeepSeek-R1-Distill-Qwen-32B Gold evidence 0.17
- 9
DeepSeek-R1-Distill-Qwen-14B Gold evidence 0.16
- 10
DeepSeek-R1-Distill-Qwen-7B Top 20 paragraphs 0.16
- 11
Gemma 3 27B Full text 0.16
- 12 GPT-3.5 Turbo Jun 2023, Gold evidence 0.15
- 13
DeepSeek-R1-Distill-Llama-70B Gold evidence 0.15
- 14 Command-R v01-34B-128k Gold evidence 0.13
Source: The PeerQA paper
- 1
Mistral-7B-Instruct-v0.2 Gold evidence 0.24
- 2 GPT-4o Aug 2024, Gold evidence 0.20
- 3
Llama 3.3 70B Instruct Full text 0.18
- 4 deepseek-r1-0528-qwen3-8b Gold evidence 0.17
- 5
DeepSeek-R1-Distill-Qwen-32B Gold evidence 0.17
- 6
DeepSeek-R1-Distill-Qwen-14B Gold evidence 0.16
- 7
DeepSeek-R1-Distill-Qwen-7B Top 20 paragraphs 0.16
- 8
Gemma 3 27B Full text 0.16
- 9 GPT-3.5 Turbo Jun 2023, Gold evidence 0.15
- 10
DeepSeek-R1-Distill-Llama-70B Gold evidence 0.15
Source: The PeerQA paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 0.24 | Gold evidence | Dec 2023 | Official leaderboard |
| 2 | GPT-4o Aug 2024 | 0.20 | Gold evidence | Aug 2024 | Official leaderboard |
| 3 | Llama-3 IT-8B-32k | 0.18 | Full text | Official leaderboard | |
| 4 | | 0.18 | Full text | Dec 2024 | Official leaderboard |
| 5 | Llama-3 IT-8B-8k | 0.18 | Top 20 paragraphs | Official leaderboard | |
| 6 | deepseek-r1-0528-qwen3-8b | 0.17 | Gold evidence | May 2025 | Official leaderboard |
| 7 | DeepSeek-R1-Llama-8B-128k | 0.17 | Gold evidence | Official leaderboard | |
| 8 | | 0.17 | Gold evidence | Jan 2025 | Official leaderboard |
| 9 | | 0.16 | Gold evidence | Jan 2025 | Official leaderboard |
| 10 | | 0.16 | Top 20 paragraphs | Jan 2025 | Official leaderboard |
| 11 | | 0.16 | Full text | Mar 2025 | Official leaderboard |
| 12 | GPT-3.5 Turbo Jun 2023 | 0.15 | Gold evidence | Jun 2023 | Official leaderboard |
| 13 | | 0.15 | Gold evidence | Jan 2025 | Official leaderboard |
| 14 | Command-R v01-34B-128k | 0.13 | Gold evidence | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 0.24 | Gold evidence | Dec 2023 | Official leaderboard |
| 2 | GPT-4o Aug 2024 | 0.20 | Gold evidence | Aug 2024 | Official leaderboard |
| 3 | | 0.18 | Full text | Dec 2024 | Official leaderboard |
| 4 | deepseek-r1-0528-qwen3-8b | 0.17 | Gold evidence | May 2025 | Official leaderboard |
| 5 | | 0.17 | Gold evidence | Jan 2025 | Official leaderboard |
| 6 | | 0.16 | Gold evidence | Jan 2025 | Official leaderboard |
| 7 | | 0.16 | Top 20 paragraphs | Jan 2025 | Official leaderboard |
| 8 | | 0.16 | Full text | Mar 2025 | Official leaderboard |
| 9 | GPT-3.5 Turbo Jun 2023 | 0.15 | Gold evidence | Jun 2023 | Official leaderboard |
| 10 | | 0.15 | Gold evidence | Jan 2025 | Official leaderboard |
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: The PeerQA paper
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: The PeerQA paper
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The PeerQA paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official PeerQA leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Science and reasoning
- GPQA Diamond In use 198 graduate-level 'Google-proof' multiple-choice questions in biology, chemistry and physics.
Gemini 4 Carbon 97%
- Terminal-Bench-Science In use 70 scientific research-workflow tasks across five science areas, run in Harbor containers. Hidden tests give the resolution rate. GPT-6 Astra 68.1%
- BioMysteryBench In use Bioinformatics problems on anonymized real datasets that require real analysis with tools. Rubric-graded, with human-solvable and human-difficult splits. Claude Fable 5.2 83.7%
- FrontierScience In use Models solve physics, chemistry, and biology problems in an Olympiad track graded by short answer and a Research track graded by 10-point rubrics.
Intern-S2-397B 87%
- HiPhO In use Models solve theory problems from 2024 and 2025 physics olympiad exams, some with diagrams. Official marking schemes grade answers and steps, mapped to medal levels.
Gemini 3 Pro 78.3%
- LABBench2 In use 11 biology research sub-tasks (literature, figures, tables, patents, trials and more) with tool access. Macro-averaged accuracy. GPT-5.2 60.9%
Data audit
8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 18 months ago, in May 2025, for deepseek-r1-0528-qwen3-8b. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
- Pass Independent runs Scores from the benchmark’s paper (14 entries).
- Pass Models scored 10 models we track, plus 4 other entries on its leaderboard.
- Warning Newest score The newest came 18 months ago, in May 2025, for deepseek-r1-0528-qwen3-8b.
- Note Frontier models None of the top 10 has a score. The strongest tested is Gemma 3 27B (#165).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on github.com: 51 rows for models we track and 4 other entries at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.