BenchCAD 2.0
Models reconstruct parts, assemblies, and circuit boards from drawings and images. The Core set contains 100 problems with partial credit.
Models use 30 rounds with compaction at 180,000 tokens. Scores do not compare with 2.1-beta or 2.1. Core requires a Hugging Face account with access. Public samples are available.
- Made by
- BenchCAD team
- Released
- October 2026
- Kind
- CAD models
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- Public, plus a private test set
- Strongest model tested
- GPT-6 Astra #3 on the capability index
- Last model added
- October 2026
- Scores refreshed
- November 12th, 2026
OpenAI internal model (unreleased) tops the leaderboard with 87.5%
Best score per entry from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. 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 GPT-6 Astra, at 59%.
- 1 OpenAI internal model (unreleased) 87.5%
- 2 GPT-6 Astra max 59%
- 3 Claude Opus 5.5 max 57%
- 4 Claude Sonnet 5.5 max 41%
- 5 Claude Fable 5.1 max 30%
- 6 Claude Haiku 5.5 max 18%
Source: Official BenchCAD 2.0 leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | OpenAI internal model (unreleased) | 87.5% | Codex | Official leaderboard | |
| 2 | GPT-6 Astra | 59% | max, BenchCAD 2.0 | Sep 2026 | Official leaderboard |
| 3 | Claude Opus 5.5 | 57% | max, BenchCAD 2.0 | Sep 2026 | Official leaderboard |
| 4 | Claude Sonnet 5.5 | 41% | max, BenchCAD 2.0 | Sep 2026 | Official leaderboard |
| 5 | Claude Fable 5.1 | 30% | max, BenchCAD 2.0 | Sep 2026 | Official leaderboard |
| 6 | Claude Haiku 5.5 | 18% | max, BenchCAD 2.0 | Oct 2026 | Official leaderboard |
The best score, 59%, costs $11 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.
Source: Official BenchCAD 2.0 leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| GPT-6 Astra | max | BenchCAD 2.0 | 59% | $11 | |
| Claude Opus 5.5 | max | BenchCAD 2.0 | 57% | $28 | |
| GPT-6 Astra | xhigh | BenchCAD 2.0 | 45% | $7.00 | |
| GPT-6 Astra | high | BenchCAD 2.0 | 41% | $5.00 | |
| Claude Sonnet 5.5 | max | BenchCAD 2.0 | 41% | $43 | |
| Claude Opus 5.5 | xhigh | BenchCAD 2.0 | 32% | $8.00 | |
| GPT-6 Astra | medium | BenchCAD 2.0 | 31% | $3.00 | |
| Claude Sonnet 5.5 | xhigh | BenchCAD 2.0 | 31% | $5.00 | |
| Claude Fable 5.1 | max | BenchCAD 2.0 | 30% | $48 | |
| Claude Fable 5.1 | xhigh | BenchCAD 2.0 | 27% | $32 | |
| Claude Opus 5.5 | high | BenchCAD 2.0 | 18% | $2.00 | |
| Claude Haiku 5.5 | max | BenchCAD 2.0 | 18% | $9.00 | |
| Claude Haiku 5.5 | xhigh | BenchCAD 2.0 | 15% | $0.81 | |
| GPT-6 Astra | low | BenchCAD 2.0 | 14% | $0.44 | |
| Claude Sonnet 5.5 | high | BenchCAD 2.0 | 14% | $1.00 | |
| Claude Fable 5.1 | high | BenchCAD 2.0 | 13% | $11 | |
| Claude Opus 5.5 | medium | BenchCAD 2.0 | 13% | $0.68 | |
| Claude Fable 5.1 | medium | BenchCAD 2.0 | 11% | $5.00 | |
| Claude Fable 5.1 | low | BenchCAD 2.0 | 9.00% | $2.00 | |
| Claude Haiku 5.5 | high | BenchCAD 2.0 | 8.00% | $0.06 | |
| Claude Opus 5.5 | low | BenchCAD 2.0 | 6.00% | $0.06 | |
| Claude Sonnet 5.5 | medium | BenchCAD 2.0 | 6.00% | $0.15 | |
| Claude Sonnet 5.5 | low | BenchCAD 2.0 | 5.00% | $0.05 | |
| Claude Haiku 5.5 | medium | BenchCAD 2.0 | 3.00% | $0.03 | |
| Claude Haiku 5.5 | low | BenchCAD 2.0 | 1.00% | $0.02 |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official BenchCAD 2.0 leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official BenchCAD 2.0 leaderboard
No lab model card reports it yet
We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.
Versions
We switch the top row to each new version when it comes out. Open it for one row per version. An old version stays in play while people still test it, so whoever first takes it past 85% gets credit for closing it out.
Sources: Official leaderboards and labs’ model cards
More 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% - Parametric CAD Bench In use Agents author editable parametric FreeCAD part designs, graded against held-back references for geometry and spec consistency. Claude Opus 5.5 61%
- 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%
- LMBuild Not in model cards Language model agents design 3D structures as parts, joints, materials and assembly steps. Scoring checks that the structures can be built and work as intended. GPT-6 Astra 66.5%