Benchmark Almanac

SuperCLUE-Science

o4-mini 80.8 leads since April 2025
Not in model cards

Models solve Chinese science reasoning problems.

SuperCLUE reports scores in a public workbook.

Made by
SuperCLUE
Kind
Science questions
Level
University
Scored
Checked automatically
Answers
Short answer
Measures
Capability
Language
Chinese
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
A host
Strongest model tested
o4-mini #95 on the capability index
Last model added
April 2025
Scores refreshed
November 20th, 2026
23
Entries on its own leaderboard, 16 models we track

o4-mini leads with 80.8 among models released by April 2025 o4-mini leads with 80.8 among models released by April 2025

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. 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 April 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 o4-mini high / API 80.8
  2. 2 o3-mini high / API 79.1
  3. 3 Gemini 2.5 Pro Mar 2025, API 77.2
  4. 4 GPT-4.1 API 74.4
  5. 5 Qwen3-235B-A22B API 74.1
  6. 6 Claude 3.7 Sonnet POE 73.7
  7. 7 DeepSeek-R1 API 73.6
  8. 8 Qwen3-32B API 73.5
  9. 9 o1 官网 72.6
  10. 10 DeepSeek-V3 Mar 2025, API 72.5
  11. 11 QwQ-Max-Preview 官网 72.2
  12. 12 Qwen3-30B-A3B API 69.6
  13. 13 chatgpt-4o-latest API 69.6
  14. 14 Gemini 2.0 Flash Dec 2024, API 69.2
  15. 15 QwQ-32B API 68.7
  16. 16 GLM-Zero-Preview API 65.7
  17. 17 Llama 4 Maverick API 64.9
  18. 18 Claude-3.5-Sonnet(1022) API 63.6
  19. 19 DeepSeek-R1-Lite-Preview 官网 62.7
  20. 20 Kimi-k1 官网 61.0
  21. 21 Llama 4 Scout API 61.0
  22. 22 InternThinker 官网 60.2
  23. 23 QwQ-32B-Preview API 58.1

Source: Official SuperCLUE-Science leaderboard

#EntryScoreSettingReleasedRun onSource
1 o4-mini 80.8 high / API Apr 2025 Apr 2025 Official leaderboard
2 o3-mini 79.1 high / API Jan 2025 Mar 2025 Official leaderboard
3 Gemini 2.5 Pro Mar 2025 77.2 API Mar 2025 Mar 2025 Official leaderboard
4 GPT-4.1 74.4 API Apr 2025 Apr 2025 Official leaderboard
5 Qwen3-235B-A22B 74.1 API Apr 2025 Apr 2025 Official leaderboard
6 Claude 3.7 Sonnet 73.7 POE Feb 2025 Feb 2025 Official leaderboard
7 DeepSeek-R1 73.6 API Jan 2025 Jan 2025 Official leaderboard
8 Qwen3-32B 73.5 API Apr 2025 Apr 2025 Official leaderboard
9 o1 72.6 官网 Dec 2024 Jan 2025 Official leaderboard
10 DeepSeek-V3 Mar 2025 72.5 API Mar 2025 Mar 2025 Official leaderboard
11 QwQ-Max-Preview 72.2 官网 Feb 2025 Official leaderboard
12 Qwen3-30B-A3B 69.6 API Apr 2025 Apr 2025 Official leaderboard
13 chatgpt-4o-latest 69.6 API Jan 2025 Official leaderboard
14 Gemini 2.0 Flash Dec 2024 69.2 API Dec 2024 Jan 2025 Official leaderboard
15 QwQ-32B 68.7 API Mar 2025 Mar 2025 Official leaderboard
16 GLM-Zero-Preview 65.7 API Jan 2025 Official leaderboard
17 Llama 4 Maverick 64.9 API Apr 2025 Apr 2025 Official leaderboard
18 Claude-3.5-Sonnet(1022) 63.6 API Jan 2025 Official leaderboard
19 DeepSeek-R1-Lite-Preview 62.7 官网 Jan 2025 Official leaderboard
20 Kimi-k1 61.0 官网 Jan 2025 Official leaderboard
21 Llama 4 Scout 61.0 API Apr 2025 Apr 2025 Official leaderboard
22 InternThinker 60.2 官网 Jan 2025 Official leaderboard
23 QwQ-32B-Preview 58.1 API Nov 2024 Jan 2025 Official leaderboard
#ModelScoreSettingReleasedRun onSource
1 o4-mini 80.8 high / API Apr 2025 Apr 2025 Official leaderboard
2 o3-mini 79.1 high / API Jan 2025 Mar 2025 Official leaderboard
3 Gemini 2.5 Pro Mar 2025 77.2 API Mar 2025 Mar 2025 Official leaderboard
4 GPT-4.1 74.4 API Apr 2025 Apr 2025 Official leaderboard
5 Qwen3-235B-A22B 74.1 API Apr 2025 Apr 2025 Official leaderboard
6 Claude 3.7 Sonnet 73.7 POE Feb 2025 Feb 2025 Official leaderboard
7 DeepSeek-R1 73.6 API Jan 2025 Jan 2025 Official leaderboard
8 Qwen3-32B 73.5 API Apr 2025 Apr 2025 Official leaderboard
9 o1 72.6 官网 Dec 2024 Jan 2025 Official leaderboard
10 DeepSeek-V3 Mar 2025 72.5 API Mar 2025 Mar 2025 Official leaderboard
11 Qwen3-30B-A3B 69.6 API Apr 2025 Apr 2025 Official leaderboard
12 Gemini 2.0 Flash Dec 2024 69.2 API Dec 2024 Jan 2025 Official leaderboard
13 QwQ-32B 68.7 API Mar 2025 Mar 2025 Official leaderboard
14 Llama 4 Maverick 64.9 API Apr 2025 Apr 2025 Official leaderboard
15 Llama 4 Scout 61.0 API Apr 2025 Apr 2025 Official leaderboard
16 QwQ-32B-Preview 58.1 API Nov 2024 Jan 2025 Official leaderboard

Rankings by part

We read 3 parts from the benchmark’s own leaderboard: 物理, 化学 and 生物. Each ranks models on one part of the test. o4-mini leads 2 of them, and Gemini 2.5 Pro leads the other one. Arrows show how far a model moved from its overall rank among the models we track.

物理

16 models we track have a score here.

  1. 1 o4-mini 86.2
  2. 2 o3-mini 85.0
  3. 3 Qwen3-235B-A22B ↑2 82.7
  4. 4 Claude 3.7 Sonnet ↑2 81.1
  5. 5 Qwen3-32B ↑3 80.7
  6. 6 Gemini 2.5 Pro Mar 2025 ↓3 79.7
  7. 7 GPT-4.1 ↓3 79.4
  8. 8 Qwen3-30B-A3B ↑3 78.4
  9. 9 DeepSeek-R1 ↓2 78.0
  10. 10 DeepSeek-V3 Mar 2025 75.3

6 models have held number one on the official leaderboard since 2023

Lab claims count from the day the card came out, independent runs from the model’s release.

Qwen Google DeepMind OpenAI DeepSeek
Own leaderboard Not tested since Apr 2025
o4-mini 80.8

Source: Official SuperCLUE-Science leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard
60.0 70.0 80.0 2025 60.0 70.0 80.0 2025

Source: Official SuperCLUE-Science 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

All 203 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 official leaderboard (23 entries).
  • Pass Models scored 16 models we track, plus 7 other entries on its leaderboard.
  • Warning Newest score The newest came 19 months ago, in April 2025, for Qwen3-235B-A22B.
  • Note Frontier models None of the top 10 has a score. The strongest tested is o4-mini (#95).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on superclueai.com: 16 rows for models we track and 7 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.

How we check