Which models can actually do CAD?

PartForge is an agentic coding harness for parametric CAD. An AI agent works with you to turn a prompt, and whatever guidance follows it, into a small program that generates a 3D model. Along the way it has the documentation, a reference library of published parts, and web search. After each step of a build it gets screenshots and measurements of the thing it has made. That makes this an unusually direct way to measure how the current models handle vision, tool use and reasoning about 3D geometry at once.

This run is the hard end of the suite only: 7 models over the 27 hardest tasks, 567 attempts in all, up to three per task for each model. Each attempt is graded on whether it built, on measured geometry, and on a checklist an AI judge answers from the renders, and a person ranked every task by hand. Click any number to open the attempt behind it. The first baseline, over the whole suite from easy to hard, is still on its own page.

Quality

against

Gemini 4 Argon leads on human rank. Haiku 5.5 is the cheapest to run. No model is both. 100% is first on every task, 0% is last, 50% is tied with the field.

What we found

Cost and time still spread much further than quality did. Among the models that finished every task, the field runs from $0.02 to $0.38 a task, and from 48s to 3m 46s.

The human ranking put Gemini 4 Argon first, only just ahead of Gemini 3.8 Flash. It costs 17% less than Opus 5.5 does.

The two cheapest models, Haiku 5.5 and Sonnet 5.5, cost $0.13 a task or less — and Sonnet 5.5 has the top score of the run.

Where the score and the reviewer disagree

Overall gives partial credit at every layer, so once every model can do most of a task the scores bunch: this run's seven sit inside 4.1 points. Where the two orders below cross, the score is missing something the reviewer saw.

All models

Rank by

#	Model	Overall	Strict	Human rank	Gate pass
1	Sonnet 5.5Anthropic	91	63%	49%	96%
2	GPT-6 SolOpenAI	90	53%	44%	93%
3	GPT-6.1 SolOpenAI	88	63%	41%	93%
4	Opus 5.5Anthropic	88	54%	54%	93%
5	Gemini 3.8 FlashGoogle	88	67%	60%	93%
6	Haiku 5.5Anthropic	88	57%	41%	91%
7	Gemini 4 ArgonGoogle	87	68%	61%	96%

Overall gives partial credit at every layer, so the top of the board bunches. Strict counts only the attempts that got everything right. Gate pass is the share whose part was watertight, hole-correct and overlap-free. Hover any bar for its number. Ties break by cost, then time. In Quality, the bar marked with a person is one reviewer's ranking of the cases they ordered by hand. First place is 100%, last is 0%, ties share, and the tick marks 50%. That reviewer marked some first attempts unacceptable: Gemini 3.8 Flash 96%, Haiku 5.5 96%. Neither affects Overall or Strict.

These are the suite's hardest tasks: parts that have to fit a bearing, a motor or a Raspberry Pi, rebuilds from a photograph or a scanned mesh, and edits made over several turns to a part that already exists. Every one of them grades what the model built rather than what it said about it. The methodology page has the details.

Browse the tasks

How the score is made

Built 30%, measured geometry 50%, AI judge 20%. A failed gate zeroes geometry for that attempt. The AI judges are Opus 5.5 and GPT-6 Sol. Read the methodology.