SimLife-BP (Vision + Language) - No Hints: leaderboard
SimLife-BP long-horizon pattern understanding over simulated household life (381 tasks / 1,439 questions), vision plus language input, asked with no hints. Question-level accuracy; the input modality is the board. Higher is better. Source: arxiv.org/abs/2609.19610.
Metric: Question-level Accuracy (%). Source: arxiv.org. Saturation forecast: Around 2029. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash (Preview) | 54.8 | #90 |
| 2 | Qwen 3 VL 32B Instruct | 53 | #290 |
| 3 | Qwen 3.5 35B A3B | 52.5 | #268 |
| 4 | Qwen 3.5 9B | 51 | #391 |
| 5 | Qwen 3 VL 4B Instruct | 50.9 | #510 |
| 6 | Qwen 3 VL 8B Instruct | 49.3 | #419 |
| 7 | Gemma 4 26B A4B (IT) | 49.3 | #293 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=simlife-bp-vision-plus-language-no-hints · How It Works · Data refreshed daily, snapshot 2026-11-05.