BIBIMBAP
Agents read and edit IFC building files. Executable tests compare their changes and answers with reference results.
The 100 tasks cover five categories. The authors compared six models using IfcOpenShell and repeated the three highest scoring models three times.
- Released
- March 2026
- Kind
- CAD models
- Level
- University
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Subject
- Civil engineering
- Language
- English
- Interaction
- Agentic
- Environment
- IfcOpenShell
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Agents, Image, video and 3D generation
- Strongest model tested
- GPT-5.4 #28 on the capability index
- Last model added
- March 2026
- Scores refreshed
- November 20th, 2026
Claude Opus 4.6 leads with 50.2% among models released by March 2026
Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best.
The newest model with a score here came out in March 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1 Claude Opus 4.6 50.2%
- 2 GPT-5.4 31.3%
- 3
Gemini 3 Flash 30.8%
- 4 GPT-5.2 30.5%
- 5 GPT-5 mini 25%
- 6 GPT-4.1 23.9%
Source: Official BIBIMBAP leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.6 | 50.2% | Feb 2026 | Official leaderboard | |
| 2 | GPT-5.4 | 31.3% | Mar 2026 | Official leaderboard | |
| 3 | | 30.8% | Dec 2025 | Official leaderboard | |
| 4 | GPT-5.2 | 30.5% | Dec 2025 | Official leaderboard | |
| 5 | GPT-5 mini | 25% | Aug 2025 | Official leaderboard | |
| 6 | GPT-4.1 | 23.9% | Apr 2025 | Official leaderboard |
Rankings by part
We read 9 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Direct, Spatial and Topological. Claude Opus 4.6 leads 8 of them, and GPT-5 mini leads the other one. Arrows show how far a model moved from its overall rank among the models we track.
Direct
6 models we track have a score here.
- 1 Claude Opus 4.6 65.8%
- 2
Gemini 3 Flash ↑1 65.4%
- 3 GPT-5.4 ↓1 62.1%
- 4 GPT-5.2 52.5%
- 5 GPT-5 mini 50.9%
- 6 GPT-4.1 38.3%
Spatial
6 models we track have a score here.
- 1 Claude Opus 4.6 63.9%
- 2 GPT-5.4 42.1%
- 3
Gemini 3 Flash 41%
- 4 GPT-5.2 34.6%
- 5 GPT-4.1 ↑1 30%
- 6 GPT-5 mini ↓1 25.8%
Topological
6 models we track have a score here.
- 1 Claude Opus 4.6 53.1%
- 2 GPT-5.2 ↑2 35.4%
- 3 GPT-4.1 ↑3 31.3%
- 4 GPT-5 mini ↑1 25.8%
- 5
Gemini 3 Flash ↓2 21.1%
- 6 GPT-5.4 ↓4 20.6%
Numeric
6 models we track have a score here.
- 1 Claude Opus 4.6 38.3%
- 2 GPT-5.4 26.7%
- 3
Gemini 3 Flash 20%
- 4 GPT-5.2 5.00%
- 5 GPT-4.1 ↑1 0%
- 6 GPT-5 mini ↓1 0%
Conceptual
6 models we track have a score here.
- 1 Claude Opus 4.6 30%
- 2 GPT-5 mini ↑3 25%
- 3 GPT-5.2 ↑1 25%
- 4 GPT-4.1 ↑2 20%
- 5
Gemini 3 Flash ↓2 6.67%
- 6 GPT-5.4 ↓4 5.00%
Create
6 models we track have a score here.
- 1 Claude Opus 4.6 81.1%
- 2 GPT-5.2 ↑2 58.9%
- 3 GPT-5.4 ↓1 56%
- 4
Gemini 3 Flash ↓1 45.6%
- 5 GPT-5 mini 31.2%
- 6 GPT-4.1 29.4%
Read
6 models we track have a score here.
- 1 Claude Opus 4.6 47.3%
- 2 GPT-5.4 22.4%
- 3
Gemini 3 Flash 20.6%
- 4 GPT-5.2 18.2%
- 5 GPT-5 mini 10.9%
- 6 GPT-4.1 9.00%
Update
6 models we track have a score here.
- 1 Claude Opus 4.6 61.5%
- 2 GPT-4.1 ↑4 60%
- 3 GPT-5 mini ↑2 55.5%
- 4 GPT-5.2 51.1%
- 5 GPT-5.4 ↓3 50.4%
- 6
Gemini 3 Flash ↓3 46.7%
Delete
6 models we track have a score here.
- 1 GPT-5 mini ↑4 43.3%
- 2
Gemini 3 Flash ↑1 37.8%
- 3 GPT-4.1 ↑3 36.6%
- 4 GPT-5.2 26.7%
- 5 GPT-5.4 ↓3 20%
- 6 Claude Opus 4.6 ↓5 18.9%
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official BIBIMBAP 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 BIBIMBAP leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
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%
- 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%
- Razavi-Bench Not in model cards Models explain what happens in analog circuits and why. A language model awards up to four points per answer against a reference and a rubric.
Gemini 4 Carbon 98.3%
- CircuitRubric Not in model cards Models write analog circuit netlists from design requests. A program compares transistor connections and relative device sizes with accepted reference circuits. Claude Opus 4.8 96%
Data audit
9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 8 months ago, in March 2026, for GPT-5.4. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
- Pass Independent runs Scores from the official leaderboard (6 entries).
- Pass Models scored 6 models we track.
- Warning Newest score The newest came 8 months ago, in March 2026, for GPT-5.4.
- Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.4 (#28).
- Pass Headroom The best is 50.2%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on nbharathik.github.io: 6 rows for models we track at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.