ChemInsight
Models solve related chemistry problems in bundles. The authors score how many parts a model can solve and how far its reasoning remains correct.
- Made by
- Changyu Hu and colleagues
- Released
- November 2026
- Kind
- Science questions
- Level
- University
- Scored
- Graded by a model
- Answers
- Open-ended
- Measures
- Capability
- Subject
- Chemical sciences
- Language
- English
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- A host
- In views
- Chemistry and materials
- Strongest model tested
-
Gemini 3 Pro #56 on the capability index
- Last model added
- December 2025
- Scores refreshed
- November 20th, 2026
Gemini 3 Pro leads with 0.28 among models released by December 2025 Gemini 3 Pro leads with 0.28 among models released by December 2025
Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. 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 December 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
Gemini 3 Pro Thinking 0.28
- 2 GPT-5 Thinking 0.21
- 3 Gemini 2.5 Pro (snapshot unspecified) Thinking 0.21
- 4
DeepSeek-V3.2-Speciale Thinking 0.21
- 5
DeepSeek-V3.2 Thinking 0.18
- 6
Qwen3-235B-A22B-Thinking Jul 2025, Thinking 0.15
- 7
Ring-1T Thinking 0.15
- 8 DeepSeek-R1 (snapshot unspecified) Thinking 0.14
- 9 Ling-1T Non-thinking 0.12
- 10
DeepSeek-V3.1 Non-thinking 0.12
Source: The ChemInsight paper
- 1
Gemini 3 Pro Thinking 0.28
- 2 GPT-5 Thinking 0.21
- 3
DeepSeek-V3.2-Speciale Thinking 0.21
- 4
DeepSeek-V3.2 Thinking 0.18
- 5
Qwen3-235B-A22B-Thinking Jul 2025, Thinking 0.15
- 6
Ring-1T Thinking 0.15
- 7
DeepSeek-V3.1 Non-thinking 0.12
Source: The ChemInsight paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 0.28 | Thinking | Nov 2025 | Official leaderboard |
| 2 | GPT-5 | 0.21 | Thinking | Aug 2025 | Official leaderboard |
| 3 | Gemini 2.5 Pro (snapshot unspecified) | 0.21 | Thinking | Official leaderboard | |
| 4 | | 0.21 | Thinking | Dec 2025 | Official leaderboard |
| 5 | | 0.18 | Thinking | Dec 2025 | Official leaderboard |
| 6 | | 0.15 | Thinking | Jul 2025 | Official leaderboard |
| 7 | | 0.15 | Thinking | Oct 2025 | Official leaderboard |
| 8 | DeepSeek-R1 (snapshot unspecified) | 0.14 | Thinking | Official leaderboard | |
| 9 | Ling-1T | 0.12 | Non-thinking | Official leaderboard | |
| 10 | | 0.12 | Non-thinking | Aug 2025 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 0.28 | Thinking | Nov 2025 | Official leaderboard |
| 2 | GPT-5 | 0.21 | Thinking | Aug 2025 | Official leaderboard |
| 3 | | 0.21 | Thinking | Dec 2025 | Official leaderboard |
| 4 | | 0.18 | Thinking | Dec 2025 | Official leaderboard |
| 5 | | 0.15 | Thinking | Jul 2025 | Official leaderboard |
| 6 | | 0.15 | Thinking | Oct 2025 | Official leaderboard |
| 7 | | 0.12 | Non-thinking | Aug 2025 | Official leaderboard |
Rankings by part
We read 6 parts from the benchmark’s paper, each ranking models on one part of the test, such as Reaction-type recognition, Structure generation in context and Calculation and counting. Gemini 3 Pro leads 3 of them, and GPT-5 leads the other 3. Arrows show how far a model moved from its overall rank among the models we track.
Reaction-type recognition
7 models we track have a score here.
- 1
Gemini 3 Pro 0.22
- 2
DeepSeek-V3.2-Speciale ↑1 0.12
- 3
DeepSeek-V3.2 ↑1 0.09
- 4 GPT-5 ↓2 0.08
- 5
Qwen3-235B-A22B-Thinking Jul 2025 0.06
- 6
Ring-1T 0.05
- 7
DeepSeek-V3.1 0.01
Structure generation in context
7 models we track have a score here.
- 1
Gemini 3 Pro 0.20
- 2
DeepSeek-V3.2-Speciale ↑1 0.12
- 3
DeepSeek-V3.2 ↑1 0.09
- 4 GPT-5 ↓2 0.08
- 5
Qwen3-235B-A22B-Thinking Jul 2025 0.06
- 6
Ring-1T 0.05
- 7
DeepSeek-V3.1 0.01
Calculation and counting
7 models we track have a score here.
- 1
Gemini 3 Pro 0.55
- 2
DeepSeek-V3.2-Speciale ↑1 0.37
- 3
DeepSeek-V3.2 ↑1 0.36
- 4 GPT-5 ↓2 0.28
- 5
Ring-1T ↑1 0.26
- 6
DeepSeek-V3.1 ↑1 0.24
- 7
Qwen3-235B-A22B-Thinking Jul 2025 ↓2 0.12
Structure-property reasoning
7 models we track have a score here.
- 1 GPT-5 ↑1 0.45
- 2
Gemini 3 Pro ↓1 0.40
- 3
DeepSeek-V3.2-Speciale 0.35
- 4
DeepSeek-V3.2 0.35
- 5
Qwen3-235B-A22B-Thinking Jul 2025 0.33
- 6
Ring-1T 0.32
- 7
DeepSeek-V3.1 0.30
Selectivity prediction
7 models we track have a score here.
- 1 GPT-5 ↑1 0.36
- 2
Gemini 3 Pro ↓1 0.29
- 3
DeepSeek-V3.2-Speciale 0.28
- 4
Ring-1T ↑2 0.28
- 5
Qwen3-235B-A22B-Thinking Jul 2025 0.26
- 6
DeepSeek-V3.2 ↓2 0.24
- 7
DeepSeek-V3.1 0.13
Spatial imagination and three-dimensional thinking
7 models we track have a score here.
- 1 GPT-5 ↑1 0.21
- 2
Gemini 3 Pro ↓1 0.19
- 3
DeepSeek-V3.2-Speciale 0.16
- 4
DeepSeek-V3.1 ↑3 0.16
- 5
DeepSeek-V3.2 ↓1 0.16
- 6
Ring-1T 0.14
- 7
Qwen3-235B-A22B-Thinking Jul 2025 ↓2 0.13
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The ChemInsight 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 ChemInsight 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
7 of 9 checks pass, 2 warnings. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark
Catalog record 1 warning 1 pass
- Warning Task access We don’t know whether the 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 (10 entries).
- Pass Models scored 7 models we track, plus 3 other entries on its leaderboard.
- Warning Newest score The newest came 11 months ago, in December 2025, for DeepSeek-V3.2.
- Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 3 Pro (#56).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 7 rows for models we track and 3 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.