partforge
Browse forgesSign in

Which models can actually do CAD?

7 models, 27 tasks, 567 attempts. Each attempt is graded on whether it built, on measured geometry, and on a checklist an AI judge answers from the renders. Click any number to open the attempt behind it.

Quality
against
0%25%50%75%100%$0.00$0.10$0.20$0.30$0.400s50s1m 40s2m 30s3m 20s4m 10sSonnet 5.5GPT-6 SolGPT-6.1 SolOpus 5.5Gemini 3.8 FlashHaiku 5.5Gemini 4 Argon
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.

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.

Overall scoreHuman rankingSonnet 5.5: 1 by overall score, 4 by the reviewer's ranking, 3 places lowerGPT-6 Sol: 2 by overall score, 5 by the reviewer's ranking, 3 places lowerGPT-6.1 Sol: 3 by overall score, 7 by the reviewer's ranking, 4 places lowerOpus 5.5: 4 by overall score, 3 by the reviewer's ranking, 1 place higherGemini 3.8 Flash: 5 by overall score, 2 by the reviewer's ranking, 3 places higherHaiku 5.5: 6 on bothGemini 4 Argon: 7 by overall score, 1 by the reviewer's ranking, 6 places higher1Sonnet 5.52GPT-6 Sol3GPT-6.1 Sol4Opus 5.55Gemini 3.8 Flash6Haiku 5.57Gemini 4 ArgonGemini 4 Argon1Gemini 3.8 Flash2Opus 5.53Sonnet 5.54GPT-6 Sol5Haiku 5.56GPT-6.1 Sol7

All models

Rank by
#ModelQualityCost and timeOverallStrictHuman rankGate pass
1Sonnet 5.5Anthropic
9163%49%96%
2GPT-6 SolOpenAI
9053%44%93%
3GPT-6.1 SolOpenAI
8863%41%93%
4Opus 5.5Anthropic
8854%54%93%
5Gemini 3.8 FlashGoogle
8867%60%93%
6Haiku 5.5Anthropic
8857%41%91%
7Gemini 4 ArgonGoogle
8768%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.

Browse the tasks

Hard

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.

Run hard-compact-2026-10-162026-10-16commit 6fa9f854spend $116.04 + $54.69 AI judging, 1 unpricedOther runs: 2026-09-23T06-56-16-hard-baseline-2026-09-23, 2026-09-10T19-44-23-baseline-2026-09, 2026-09-04T20-02-10-stage2-hard-cases, 2026-09-04T19-48-31-screen-14-fixed