TECCI - Pairwise Elo (Instruction Rich Challenge Set): leaderboard
Side-by-side human preference between image-editing models on TECCI's manually written challenge instructions (Agrawal et al., 2026), three raters per pair. Elo rating; higher is better.
Metric: Elo rating from side-by-side human preference between two models' edits (three raters per pair), manually written instructions. Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Nano Banana 2 (Gemini 3.1 Flash Image Preview) | 1055 | |
| 2 | Nano Banana Pro | 1045 | |
| 3 | Grok Imagine Pro | 987 | |
| 4 | GPT Image 1.5 | 971 | |
| 5 | Seedream 5.0 Lite | 941 |
Interactive version: theaggregate.ai/benchmark?slug=tecci-pairwise-elo-instruction-rich-challenge-set · How It Works · Data refreshed daily, snapshot 2026-11-12.