Elliot Caption Quality - Mean rank (1–4): leaderboard
Elliot Caption Quality - Mean rank (1–4). 100 stratified tiles;blind ClaudeOpus5.5 judge sees image,fact sheet,four randomized captions. Mean over tiles; source Table5. Input conditions retained on exact run labels. Metric: Mean rank (1–4). Primary published table; source configurations remain distinct.
Metric: Mean rank (1–4). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6.1 Sol (Elliot Run: Fact sheet) | 1.55 | |
| 2 | GPT-6.1 Sol (Elliot Run: Fact sheet and image) | 1.64 | |
| 3 | GPT-6 Luna (Elliot Run: Fact sheet) | 3.49 |
Interactive version: theaggregate.ai/benchmark?slug=elliot-caption-quality-mean-rank-1-4 · How It Works · Data refreshed daily, snapshot 2026-10-16.