Benchmark Almanac

P3D-Bench

GPT-6 Astra 69.4 leads since September 2026
Not in model cards

Models generate parametric 3D objects from text, images, or assembly instructions. The evaluator checks whether outputs run and preserve geometry, topology, and part structure.

Made by
Yikang Yang and colleagues
Released
June 2026
Kind
Image generation and editing
Level
University
Scored
Checked automatically
Answers
Code
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Images → Code
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
In views
CAD, Vision
Strongest model tested
GPT-6 Astra #4 on the capability index
Last model added
September 2026
Scores refreshed
November 20th, 2026
9
Models scored on its own leaderboard

GPT-6 Astra leads with 69.4 among models released by September 2026

Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. The board reports no overall score, so the headline is the plain mean of its three task scores. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart.

The newest model with a score here came out in September 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 GPT-6 Astra 69.4
  2. 2 Gemini 3.8 Flash 63.0
  3. 3 Claude Opus 5 62.6
  4. 4 Kimi K3 61.2
  5. 5 Grok 4.6 60.8
  6. 6 Qwen 3.8 Max 60.8
  7. 7 GLM-5.3-Flash 58.3
  8. 8 DeepSeek V4.1 Flash 58.2
  9. 9 Doubao Seed 2.1 Pro 49.5

Source: Official P3D-Bench leaderboard

#ModelScoreSettingDateSource
1 GPT-6 Astra 69.4 Sep 2026 Official leaderboard
2 Gemini 3.8 Flash 63.0 Sep 2026 Official leaderboard
3 Claude Opus 5 62.6 Jul 2026 Official leaderboard
4 Kimi K3 61.2 Jul 2026 Official leaderboard
5 Grok 4.6 60.8 Aug 2026 Official leaderboard
6 Qwen 3.8 Max 60.8 Aug 2026 Official leaderboard
7 GLM-5.3-Flash 58.3 Aug 2026 Official leaderboard
8 DeepSeek V4.1 Flash 58.2 Sep 2026 Official leaderboard
9 Doubao Seed 2.1 Pro 49.5 Jun 2026 Official leaderboard

Rankings by part

We read 52 parts from the benchmark’s own leaderboard in 4 groups: task, assembly-3d, text-to-3d and image-to-3d. Each ranks models on one part of the test. GPT-6 Astra leads 27 of them, and 6 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Assembly-3D

9 models we track have a score here.

  1. 1 GPT-6 Astra 63.5
  2. 2 Claude Opus 5 ↑1 54.2
  3. 3 Gemini 3.8 Flash ↓1 53.2
  4. 4 Kimi K3 50.8
  5. 5 Grok 4.6 50.7
  6. 6 Qwen 3.8 Max 49.7
  7. 7 GLM-5.3-Flash 48.8
  8. 8 DeepSeek V4.1 Flash 47.5
  9. 9 Doubao Seed 2.1 Pro 29.7

The best score, 69.4, costs $0.89 a task

Runs on the cost frontier score higher than any run that's cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

Qwen Z.ai Google DeepMind OpenAI Cost frontier One model at each effort level
0.00 20.0 40.0 60.0 80.0 $0.01 $0.02 $0.05 $0.10 $0.20 $0.50 $1.00 Cost per task, log scale GPT-6 Astra Gemini 3.8 Flash GLM-5.3-Flash 0.00 20.0 40.0 60.0 80.0 $0.01 $0.10 $1.00 Cost per task, log scale GPT-6 Astra Gemini 3.8 Flash

Source: Official P3D-Bench leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
GPT-6 Astra 69.4 $0.89
Gemini 3.8 Flash 63.0 $0.12
Claude Opus 5 62.6 $0.59
Kimi K3 61.2 $0.38
Grok 4.6 60.8 $0.19
Qwen 3.8 Max 60.8 $0.10
GLM-5.3-Flash 58.3 $0.02
DeepSeek V4.1 Flash 58.2 $0.04
Doubao Seed 2.1 Pro 49.5 $0.22

5 models have held number one on the official leaderboard since its release in June 2026

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

ByteDance Seed Moonshot AI Anthropic Google DeepMind OpenAI
Own leaderboard Not tested since Sep 2026

Source: Official P3D-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
45.0 50.0 55.0 60.0 65.0 70.0 2026 45.0 50.0 55.0 60.0 65.0 70.0 2026

Source: Official P3D-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 Vision

All 636 in Vision
  • MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy. Qwen 3.6 Plus 86%
  • MMMU-Pro In use Harder MMMU: college-level multimodal questions with more options and vision-only inputs to remove text shortcuts. GPT-5.6 Sol 84.6%
  • CharXiv In use Chart understanding on real arXiv figures. Reasoning questions require synthesizing information from complex charts. Claude Fable 5 93.5%
  • MathVista In use Math reasoning in visual contexts (charts, geometry, puzzles) from 31 datasets. Answer accuracy. Seed2.1 Pro 90.7%
  • MATH-Vision (MathVision) In use 3,040 competition math problems with visual context. Answer accuracy. Qwen 3.8 Max 97.7%
  • MMStar In use Models answer 1,500 human-curated image questions that cannot be solved without the image, across 6 core capabilities. Scored by accuracy. Qwen 3.8 Max 85.9%

Data audit

All 9 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (9 entries).
  • Pass Models scored 9 models we track.
  • Pass Newest score The newest is from September 2026, for DeepSeek V4.1 Flash.
  • Note Frontier models Only 1 of the top 10 has a score. The strongest tested is GPT-6 Astra (#4).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on spatiaos.github.io: 9 rows for models we track 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