LLM-IOAA
Models solve 49 theory problems and eight data analysis problems from four astronomy olympiads. Two expert graders score the solutions using the official rubrics.
The international olympiad organizers published the five models’ results in their resources. Synced covered the evaluation on 36Kr.
- Made by
- The Ohio State University and collaborators
- Released
- October 2025
- Kind
- Science olympiads
- Level
- Competition
- Scored
- Graded by people
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Capability
- Subject
- Physical sciences, Mathematics
- Language
- English
- Interaction
- Single turn
- Items from
- Exams and contests
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- GPT-5 #71 on the capability index
- Last model added
- August 2025
- Scores refreshed
- November 20th, 2026
GPT-5 leads with 86.3% among models released by August 2025 GPT-5 leads with 86.3% among models released by August 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 paper reports no overall score, so this is the plain mean of the theory and data analysis exam scores. Where a model ran with several settings or harnesses, we keep its best. 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 August 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
- 1 GPT-5 86.3%
- 2 Gemini 2.5 Pro (snapshot unspecified) 80.7%
- 3 o3 72.6%
- 4 Claude Opus 4.1 59.8%
- 5 Claude Sonnet 4 54.3%
Source: The LLM-IOAA paper
- 1 GPT-5 86.3%
- 2 o3 72.6%
- 3 Claude Opus 4.1 59.8%
- 4 Claude Sonnet 4 54.3%
Source: The LLM-IOAA paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5 | 86.3% | Aug 2025 | Official leaderboard | |
| 2 | Gemini 2.5 Pro (snapshot unspecified) | 80.7% | Official leaderboard | ||
| 3 | o3 | 72.6% | Apr 2025 | Official leaderboard | |
| 4 | Claude Opus 4.1 | 59.8% | Aug 2025 | Official leaderboard | |
| 5 | Claude Sonnet 4 | 54.3% | May 2025 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5 | 86.3% | Aug 2025 | Official leaderboard | |
| 2 | o3 | 72.6% | Apr 2025 | Official leaderboard | |
| 3 | Claude Opus 4.1 | 59.8% | Aug 2025 | Official leaderboard | |
| 4 | Claude Sonnet 4 | 54.3% | May 2025 | Official leaderboard |
Rankings by part
We read 17 parts from the benchmark’s paper in 2 groups: theory exams and data analysis exams. Each ranks models on one part of the test. GPT-5 leads 15 of them, and o3 leads the other 2. Arrows show how far a model moved from its overall rank among the models we track.
Theory Exams
4 models we track have a score here.
- 1 GPT-5 84.2%
- 2 o3 77.5%
- 3 Claude Opus 4.1 64.7%
- 4 Claude Sonnet 4 60.6%
Data Analysis Exams
4 models we track have a score here.
- 1 GPT-5 88.5%
- 2 o3 67.7%
- 3 Claude Opus 4.1 54.8%
- 4 Claude Sonnet 4 47.9%
2025
4 models we track have a score here.
- 1 GPT-5 86.8%
- 2 o3 73.8%
- 3 Claude Opus 4.1 69.8%
- 4 Claude Sonnet 4 60.2%
2024
4 models we track have a score here.
- 1 GPT-5 67.5%
- 2 o3 67.3%
- 3 Claude Sonnet 4 ↑1 57.8%
- 4 Claude Opus 4.1 ↓1 57.2%
2023
4 models we track have a score here.
- 1 GPT-5 89.6%
- 2 o3 88.8%
- 3 Claude Opus 4.1 62.8%
- 4 Claude Sonnet 4 57.4%
2022
4 models we track have a score here.
- 1 GPT-5 93%
- 2 o3 80.2%
- 3 Claude Opus 4.1 68.8%
- 4 Claude Sonnet 4 67%
Easy
4 models we track have a score here.
- 1 o3 ↑1 97%
- 2 Claude Opus 4.1 ↑1 91%
- 3 GPT-5 ↓2 84.6%
- 4 Claude Sonnet 4 79.2%
Medium
4 models we track have a score here.
- 1 o3 ↑1 84.5%
- 2 GPT-5 ↓1 76.6%
- 3 Claude Sonnet 4 ↑1 62%
- 4 Claude Opus 4.1 ↓1 59.9%
Hard
4 models we track have a score here.
- 1 GPT-5 91.5%
- 2 o3 78.8%
- 3 Claude Opus 4.1 60.5%
- 4 Claude Sonnet 4 57.5%
Extra Hard
4 models we track have a score here.
- 1 GPT-5 77.2%
- 2 o3 63.1%
- 3 Claude Opus 4.1 58.9%
- 4 Claude Sonnet 4 54.7%
2025
4 models we track have a score here.
- 1 GPT-5 72.3%
- 2 o3 64.7%
- 3 Claude Opus 4.1 41%
- 4 Claude Sonnet 4 30%
2024
4 models we track have a score here.
- 1 GPT-5 91%
- 2 o3 78.3%
- 3 Claude Opus 4.1 62%
- 4 Claude Sonnet 4 50.7%
2023
4 models we track have a score here.
- 1 GPT-5 100%
- 2 Claude Opus 4.1 ↑1 71.2%
- 3 Claude Sonnet 4 ↑1 64.8%
- 4 o3 ↓2 63.2%
2022
4 models we track have a score here.
- 1 GPT-5 90.7%
- 2 o3 64.7%
- 3 Claude Sonnet 4 ↑1 46%
- 4 Claude Opus 4.1 ↓1 45%
Easy
4 models we track have a score here.
- 1 GPT-5 100%
- 2 o3 91.6%
- 3 Claude Opus 4.1 80.4%
- 4 Claude Sonnet 4 75.1%
Medium
4 models we track have a score here.
- 1 GPT-5 75.5%
- 2 o3 58.8%
- 3 Claude Opus 4.1 50%
- 4 Claude Sonnet 4 35.5%
Hard
4 models we track have a score here.
- 1 GPT-5 93.1%
- 2 o3 62.4%
- 3 Claude Opus 4.1 45.2%
- 4 Claude Sonnet 4 43.6%
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The LLM-IOAA paper
GPT-5 on the official leaderboard passed 85% in August 2025
Models released before it came out already scored that high. All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official LLM-IOAA 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
All 8 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
- Pass Independent runs Scores from the benchmark’s paper (5 entries).
- Pass Models scored 4 models we track, plus 1 other entry on its leaderboard.
- Note Newest score The newest came 15 months ago, in August 2025, for GPT-5. That’s expected for a saturated test.
- Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is GPT-5 (#71).
- Note Headroom Saturated. The best is 86.3%, past 85% since August 2025.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 4 rows for models we track and 1 other entry 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.