Furniture Assembly
The Frontier
Best score over time · one chart, every benchmark
Full rankings
22 models tested · sorted by score
| # | Model | Score |
|---|---|---|
| 1 | 76.2 | |
| 2 | 71.4 | |
| 3 | 71.4 | |
| 4 | 64.3 | |
| 5 | 60.7 | |
| 6 | 57.1 | |
| 7 | 44.0 | |
| 8 | 40.5 | |
| 9 | 38.1 | |
| 10 | 34.5 | |
| 11 | 25.0 | |
| 12 | 20.2 | |
| 13 | 17.9 | |
| 14 | 17.9 | |
| 15 | 14.3 | |
| 16 | 11.9 | |
| 17 | 10.7 | |
| 18 | 8.3 | |
| 19 | 6.0 | |
| 20 | 6.0 | |
| 21 | 4.8 | |
| 22 | 2.4 |
Score distribution
Where models cluster
Correlated benchmarks
Pearson r · original research
Benchmarks that track with Furniture Assembly
Pearson correlation across models scored on both benchmarks. Closer to 1 = strongly predictive.
Frequently asked
About Furniture Assembly
What does Furniture Assembly measure?
Furniture Assembly is a knowledge benchmark in the BenchGecko catalog. 22 AI models have been tested on it. Scores range from 2.4 to 76.2 out of 100.
Which model leads on Furniture Assembly?
Claude Opus 5.5 from Anthropic leads Furniture Assembly with a score of 76.2. The median score across 22 tested models is 22.6.
Is Furniture Assembly saturated?
No · the top score is 76.2 out of 100 (76%). There is still meaningful room for improvement on Furniture Assembly.
Does Furniture Assembly predict performance on other benchmarks?
Yes · Furniture Assembly scores correlate 0.93 with Balrog across 5 shared models. Models that do well on Furniture Assembly tend to do well on Balrog.
How often is Furniture Assembly data refreshed?
BenchGecko pulls updates daily. New model scores on Furniture Assembly appear as soon as they are published by Epoch AI or the model provider.
- Category
- Knowledge
- Max score
- 100
- Models
- 22
- Updated
- 2026-10-13
Top on Furniture Assembly
Claude Opus 5.5 · 76.2GPT-6 Astra · 71.4GPT-6.1 Sol · 71.4Claude Sonnet 5.5 · 64.3Gemini 4 Argon · 60.7More knowledge benchmarks
Same category · related evaluations