AutoCAD-Bench
We show this table for reference; we do not rank on it.
A Markov Studios computer-use benchmark that asks agents to produce 2D drawings and 3D models in AutoCAD.
Completion rate on AutoCAD-Bench — July 2026 public results
We mirror the published completion rate view for AutoCAD-Bench. GPT-5.6 Sol leads the public snapshot at 46.0%, followed by GPT-5.6 Terra (14.0%) and Claude Fable 5 (10.0%). We do not use these results to rank models overall.
GPT-5.6 Sol
OpenAI
GPT-5.6 Terra
OpenAI
Claude Fable 5
Anthropic
7 modelsAgenticCurrentDisplay onlyUpdated July 2026 public results
Completion rate table (7 models)
ScoreHow AutoCAD-Bench is shown here
BenchLM mirrors Markov Studios' AutoCAD-Bench completion table across 50 tasks: 21 2D drawings and 29 3D models. Agents operate AutoCAD 2019 on Windows through visible mouse and keyboard actions and start each task from a clean drawing.
A task counts as completed when the rubric score reaches 75. The rubric combines geometry, dimensions, annotations, and presentation, with a geometry gate. BenchLM keeps the results display only because they combine model capability with a computer-use harness and application environment.
Snapshot
The published AutoCAD-Bench snapshot places GPT-5.6 Sol first at 46.0%. The third row is 36.0 points behind. The broader top-10 range is 46.0 points, so the table still separates the published systems.
7 models have been evaluated on AutoCAD-Bench. The benchmark falls in the Agentic category. AutoCAD-Bench is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About AutoCAD-Bench
Year
2026
Tasks
21 2D drawing tasks and 29 3D modeling tasks
Format
Task completion rate at a 75-point rubric threshold
Difficulty
Professional CAD computer use
AutoCAD-Bench contains 50 tasks in AutoCAD 2019 on Windows. A rubric combines geometry, dimensions, annotations, and presentation, applies a geometry gate, and counts scores of at least 75 as completed. BenchLM keeps the results display only because model quality, the computer-use harness, and the application environment all affect the result.
Freshness and provenance
Version
AutoCAD-Bench 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does AutoCAD-Bench measure?
A Markov Studios computer-use benchmark that asks agents to produce 2D drawings and 3D models in AutoCAD.
Which model leads the published AutoCAD-Bench snapshot?
GPT-5.6 Sol currently leads the published AutoCAD-Bench snapshot with 46.0% completion rate. BenchLM shows this benchmark for display only and does not use it in overall rankings.
How many models are evaluated on AutoCAD-Bench?
The July 2026 public results snapshot contains 7 AI models.
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 2,000+ readers.
One email each week. Unsubscribe anytime.