Benchmark Almanac

CADExpert

CME-CAD 80.7% tops its leaderboard Doubao-Seed-1.6 35.6% leads the models we track
Not in model cards

Models write CadQuery programs from dimensioned engineering drawings. The graders check whether each program runs and compare its 3D shape with the reference.

The original paper compares nine pretrained vision models. A follow-up paper evaluates two IterCAD models on CADExpert using the original geometry and execution metrics.

Made by
Ke Niu and colleagues
Released
December 2025
Kind
CAD models
Level
University
Scored
Checked automatically
Answers
Code
Tasks
All public
Measures
Capability
Language
English
In and out
Images, Text → Code
Interaction
Single turn
Items from
Written by experts
Refresh
One fixed set
Run by
Anyone
In views
CAD, Vision
Strongest model tested
GPT-5 mini #92 on the capability index
Last model added
August 2025
Scores refreshed
November 20th, 2026
11
Entries in its paper, 3 models we track

CME-CAD tops the leaderboard with 80.7% Doubao-Seed-1.6 leads with 35.6% 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. 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 Doubao-Seed-1.6, at 35.6%. 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:

  1. 1 CME-CAD Fine-tuned 80.7%
  2. 2 CAD-RL Fine-tuned 71.8%
  3. 3 Qwen3-VL Pretrained 37%
  4. 4 Doubao-Seed-1.6 Pretrained 35.6%
  5. 5 GPT-5 mini Pretrained 35.1%
  6. 6 Gemini2.5 Pro (snapshot unspecified) Pretrained 30.9%
  7. 7 InternVL3 Pretrained 24.6%
  8. 8 Qwen2.5-VL Pretrained 19.2%
  9. 9 InternVL2.5 Pretrained 12%
  10. 10 Phi-3.5-vision-instruct Pretrained 3.44%
  11. 11 LLaVA-1.5 Pretrained 0.73%

Source: The CADExpert paper

#EntryScoreSettingDateSource
1 CME-CAD 80.7% Fine-tuned Official leaderboard
2 CAD-RL 71.8% Fine-tuned Official leaderboard
3 Qwen3-VL 37% Pretrained Official leaderboard
4 Doubao-Seed-1.6 35.6% Pretrained Jun 2025 Official leaderboard
5 GPT-5 mini 35.1% Pretrained Aug 2025 Official leaderboard
6 Gemini2.5 Pro (snapshot unspecified) 30.9% Pretrained Official leaderboard
7 InternVL3 24.6% Pretrained Official leaderboard
8 Qwen2.5-VL 19.2% Pretrained Official leaderboard
9 InternVL2.5 12% Pretrained Official leaderboard
10 Phi-3.5-vision-instruct 3.44% Pretrained Aug 2024 Official leaderboard
11 LLaVA-1.5 0.73% Pretrained Official leaderboard
#ModelScoreSettingDateSource
1 Doubao-Seed-1.6 35.6% Pretrained Jun 2025 Official leaderboard
2 GPT-5 mini 35.1% Pretrained Aug 2025 Official leaderboard
3 Phi-3.5-vision-instruct 3.44% Pretrained Aug 2024 Official leaderboard

Number one over time

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

ByteDance Seed
Benchmark paper Not tested since Dec 2025

Source: The CADExpert paper

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
In the benchmark’s paper Best so far
0% 20% 40% 60% 80% 100% 2025 Near the ceiling (85%) Doubao-Seed-1.6 Phi-3.5-vision-instruct 0% 20% 40% 60% 80% 100% 2025 Near the ceiling (85%) Doubao-Seed-1.6 Phi-3.5-vision-instruct

Source: Official CADExpert 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 Engineering and CAD

All 112 in Engineering and CAD
  • EEBench In use 13 electrical-engineering circuit and PCB design tasks graded by deterministic build checks, simulation and BOM constraints. No LLM judge. Gemini 4 Carbon 83.9%
  • 3DCodeBench In use Agentic procedural 3D modeling via code: author engine-ready 3D assets through software APIs. Graded for executability and shape fidelity. LongCat-2.5 55.4%
  • CADGenBench Only in system cards Agents construct CAD models from design prompts. The generation split is graded for executability and geometric correctness. Claude Fable 5 45.1%
  • PartBench Only in system cards Parametric mechanical part design and CAD-style generation, graded for geometric and manufacturing fidelity. Grok 4.6 58.5%
  • Razavi-Bench Not in model cards Models explain what happens in analog circuits and why. A language model awards up to four points per answer against a reference and a rubric. Gemini 4 Carbon 98.3%
  • CircuitRubric Not in model cards Models write analog circuit netlists from design requests. A program compares transistor connections and relative device sizes with accepted reference circuits. Claude Opus 4.8 96%

Data audit

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 15 months ago, in August 2025, for GPT-5 mini. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the benchmark’s paper (11 entries).
  • Pass Models scored 3 models we track, plus 8 other entries on its leaderboard.
  • Warning Newest score The newest came 15 months ago, in August 2025, for GPT-5 mini.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5 mini (#92).
  • Pass Headroom The best is 35.6%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 3 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