Furniture Assembly
Integrity rank #11 of 65 · 33 models scored · top score 76.19 · Claude Opus 5.5
unknown
Strongest on discrimination, weakest on contamination resistance. Caveat: the public test set does not rule out training-data contamination.
Reported scores — protocols may differ. How we compare
| # | Model | Score | Evidence | Measured |
|---|---|---|---|---|
| 1 | Claude Opus 5.5 | 76.19 | unverifiedT2 | 2026-09-22 |
| 2 | GPT-6.1 Sol | 71.43 | unverifiedT2 | 2026-09-29 |
| 3 | GPT-6 Astra | 71.43 | unverifiedT2 | 2026-09-03 |
| 4 | Claude Sonnet 5.5 | 64.29 | unverifiedT2 | 2026-09-28 |
| 5 | Gemini 4 Argon | 60.71 | unverifiedT2 | 2026-09-30 |
| 6 | Claude Fable 5.1 | 57.14 | unverifiedT2 | 2026-09-01 |
| 7 | Claude Opus 5 | 44.05 | unverifiedT2 | 2026-07-24 |
| 8 | GPT-6 Sol | 40.48 | unverifiedT2 | 2026-09-22 |
| 9 | GPT-5.6 Sol | 38.1 | unverifiedT2 | 2026-07-09 |
| 10 | GPT-5.6 Terra | 34.52 | unverifiedT2 | 2026-07-09 |
| 11 | Claude Haiku 5.5 | 25 | unverifiedT2 | 2026-10-07 |
| 12 | GPT-6 Luna | 20.24 | unverifiedT2 | 2026-09-22 |
| 13 | GPT-5.5 | 20.24 | unverifiedT2 | 2026-04-23 |
| 14 | Claude Opus 4.8 | 17.86 | unverifiedT2 | 2026-05-28 |
| 15 | GPT-5.6 Luna | 17.86 | unverifiedT2 | 2026-07-09 |
| 16 | Grok 4.6 | 14.29 | unverifiedT2 | 2026-08-12 |
| 17 | GPT-5.2 | 11.9 | unverifiedT2 | 2025-12-11 |
| 18 | GPT-5.4 | 10.71 | unverifiedT2 | 2026-03-05 |
| 19 | Claude Fable 5 | 8.33 | unverifiedT2 | 2026-06-09 |
| 20 | Kimi K3 | 5.95 | unverifiedT2 | 2026-07-16 |
| 21 | DeepSeek V4.1 Flash | 5.95 | unverifiedT2 | 2026-09-09 |
| 22 | Claude Opus 4.7 | 4.76 | unverifiedT2 | 2026-04-16 |
| 23 | Gemini 3.8 Flash | 2.38 | unverifiedT2 | 2026-09-02 |
| 24 | Kimi K2.6 | 0 | unverifiedT2 | 2026-04-20 |
| 25 | Gemini 3.1 Pro | 0 | unverifiedT2 | 2026-02-19 |
| 26 | Gemini 3.6 Flash | 0 | unverifiedT2 | 2026-07-21 |
| 27 | Gemini 3.7 Flash | 0 | unverifiedT2 | 2026-08-13 |
| 28 | Grok 4.5 | 0 | unverifiedT2 | 2026-07-08 |
| 29 | Grok 4.7 | 0 | unverifiedT2 | 2026-09-21 |
| 30 | Mistral Large 4 | 0 | unverifiedT2 | 2026-10-06 |
| 31 | Claude Opus 4.5 | 0 | unverifiedT2 | 2025-11-24 |
| 32 | Claude Opus 4.6 | 0 | unverifiedT2 | 2026-02-05 |
| 33 | Qwen3.8 Max (0902) | 0 | unverifiedT2 | 2026-09-01 |
Continue from this leaderboard into the tightest head-to-head reads.
Weighted composite. Each dimension opens to the source-backed sub-signals it is computed from.
Claims drawn from cited facts, not live model generation.
The recorded metadata does not specify what this benchmark measures.
1 cited fact
This benchmark has scores from 33 tracked models, with a recorded score range spanning 76.19.
2 cited facts
Its Benchmark Integrity Index score is 97, with grade A and rank 11 out of 65 tracked benchmarks.
4 cited facts
It is not flagged as saturated. The top recorded score is 76.19, with 23.81 points of headroom. The computed top-score cluster contains 1 models.
4 cited facts
The harness classification is consistent harness. Test-set privacy is recorded as unknown. The recorded contamination-window ratio is 0.152.
3 cited facts
Furniture Assembly: Furniture Assembly as reported in Epoch AI's Capabilities Index CSV.
Claude Opus 5.5 leads Furniture Assembly at 76.19. The full leaderboard above lists every recorded measurement, not just the headline number.
33 models have recorded scores on Furniture Assembly, spanning a score spread of 76.19.
tensor.news grades Furniture Assembly A for integrity (score 97/100), ranking #11 of 65 benchmarks we assess across discrimination, saturation, contamination resistance, harness comparability, and freshness.
No — Furniture Assembly still has headroom and continues to discriminate between models rather than bunching them at the ceiling.
Its test set is unknown, so contamination risk is on the table: a high score may partly reflect training-data overlap rather than capability.
Scores are reported under a consistent harness, so comparisons on Furniture Assembly are reasonably apples-to-apples.
Every Furniture Assembly measurement is source-backed and tagged reproduced, self-reported, or unverified; the underlying record cites epoch.ai/data/processed_data_for_eci.csv.