BlogLLM Benchmark Scores as Published, Side by Side
CADGenBench: Results as Published
Every result the catalog holds for CADGenBench: 4 models from 3 developers, read from 1 page. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.
Area Engineering and DesignMaintainer Hugging FaceData version 2026103001, as of 2026-10-30
CADGenBench judges whether AI systems can produce usable 3D mechanical parts. It has two tasks: build a valid, geometrically correct 3D model from an engineering drawing, and edit an existing STEP file to apply a requested change. Each sample gets one STEP or BREP file, and scoring runs on a leaderboard against ground truth that stays private.
CADGenBench reward
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-5.6 SolOpenAI, released 2026-07-09 | 37.1% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: xhigh |
| Claude Opus 5Anthropic, released 2026-07-24 | 36.6% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: max |
| Grok 4.7xAI, released 2026-09-21 | 44.4% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: high |
| Grok 4.6xAI, released 2026-08-12 | 40.9% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: high |
Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.