Benchmark Almanac

Civil-Bench

CivilGPT 74.7 tops its leaderboard DeepSeek-R1 May 2025 64.8 leads the models we track
Not in model cards

Models answer Chinese civil engineering licensing questions. Answer keys grade the objective questions. In the paper, GPT-4 Turbo grades written answers against individual scoring points.

OpenCompass ranks 14 models across six engineering subjects. The original study evaluated 15 models on 14,823 objective questions and 269 written questions.

Kind
Exam questions
Level
University
Scored
Checked automatically
Answers
Multiple choice
Measures
Capability
Language
Chinese
In and out
Text → Text
Interaction
Single turn
Items from
Exams and contests
Refresh
One fixed set
Run by
A host
Strongest model tested
DeepSeek-R1 May 2025 #122 on the capability index
Last model added
May 2025
Scores refreshed
November 20th, 2026
14
Entries on its own leaderboard, 6 models we track

CivilGPT tops the leaderboard with 74.7 DeepSeek-R1 leads with 64.8 among models released by May 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. The best model we track is DeepSeek-R1, at 64.8. 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:

  1. 1 CivilGPT 74.7
  2. 2 Qwen-Max 70.3
  3. 3 Hunyuan-Turbo 67.4
  4. 4 Doubao-1.5-pro-32K 65.6
  5. 5 DeepSeek-R1 May 2025 64.8
  6. 6 Qwen2.5-72B 64.4
  7. 7 MiniMax-Text-01 62.5
  8. 8 Yi-Lightning 59.6
  9. 9 GLM-4-Plus 58.9
  10. 10 GPT-4o Mar 2025 52.9
  11. 11 Moonshot-v1-128k 49.2
  12. 12 Mistral Large 2 Nov 2024 44.8
  13. 13 Spark4.0-Ultra 44.6
  14. 14 ERNIE-Speed-Pro-128K 42.6

Source: Official Civil-Bench leaderboard

#EntryScoreSettingDateSource
1 CivilGPT closed weights, 72B 74.7 Official leaderboard
2 Qwen-Max closed weights 70.3 Official leaderboard
3 Hunyuan-Turbo closed weights 67.4 Official leaderboard
4 Doubao-1.5-pro-32K closed weights 65.6 Official leaderboard
5 DeepSeek-R1 May 2025 64.8 May 2025 Official leaderboard
6 Qwen2.5-72B 64.4 Sep 2024 Official leaderboard
7 MiniMax-Text-01 62.5 Jan 2025 Official leaderboard
8 Yi-Lightning 59.6 Dec 2024 Official leaderboard
9 GLM-4-Plus closed weights, 130B 58.9 Official leaderboard
10 GPT-4o Mar 2025 52.9 Mar 2025 Official leaderboard
11 Moonshot-v1-128k closed weights 49.2 Official leaderboard
12 Mistral Large 2 Nov 2024 44.8 Nov 2024 Official leaderboard
13 Spark4.0-Ultra closed weights 44.6 Official leaderboard
14 ERNIE-Speed-Pro-128K closed weights 42.6 Official leaderboard
#ModelScoreSettingDateSource
1 DeepSeek-R1 May 2025 64.8 May 2025 Official leaderboard
2 Qwen2.5-72B 64.4 Sep 2024 Official leaderboard
3 MiniMax-Text-01 62.5 Jan 2025 Official leaderboard
4 Yi-Lightning 59.6 Dec 2024 Official leaderboard
5 GPT-4o Mar 2025 52.9 Mar 2025 Official leaderboard
6 Mistral Large 2 Nov 2024 44.8 Nov 2024 Official leaderboard

2 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 DeepSeek
Own leaderboard Not tested since May 2025

Source: Official Civil-Bench 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
40.0 45.0 50.0 55.0 60.0 65.0 2025 40.0 45.0 50.0 55.0 60.0 65.0 2025

Source: Official Civil-Bench 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 Knowledge

All 332 in Knowledge
  • MMLU-Pro In use Harder 10-option version of MMLU with reasoning-focused questions across 14 subjects. Intern-S2-397B 89.8%
  • MMLU In use 57-subject multiple-choice exam questions from elementary to professional level. Scored by accuracy. Seed1.8 92.3%
  • SuperGPQA In use 26,529 graduate-level multiple-choice questions across 285 disciplines. Accuracy. Qwen 3.6 Max (Preview) 73.9%
  • WideSearch In use 200 bilingual broad info-seeking tasks: agent collects many atomic facts into a table. Scored by success rate and row/item F1. Tencent Hy4 preview 83.9%
  • BBEH micro average In use Models solve harder replacements for each of the 23 BIG-Bench Hard tasks, 4,520 examples in total. The score is the share of all examples answered correctly. Gemma 4 31B IT 74.4%
  • HLE-Diamond In use Models answer 1,000 cleaned expert questions from Humanity's Last Exam, 500 on reasoning and 500 on knowledge, closed book. Graded on exact answers. Claude Fable 5.2 77.9%

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 (14 entries).
  • Pass Models scored 6 models we track, plus 8 other entries on its leaderboard.
  • Warning Newest score The newest came 18 months ago, in May 2025, for DeepSeek-R1.
  • Note Frontier models None of the top 10 has a score. The strongest tested is DeepSeek-R1 (#122).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on specialist.opencompass.org.cn: 6 rows for models we track and 8 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