SciImpact - Model: leaderboard

SciImpact pairs of Hugging Face models that differ in downloads, from their model cards; the model sees two artifacts from the same field and picks the one with higher impact; pairwise accuracy in percent, chance 50, higher is better.

Metric: Pairwise prediction accuracy (%, times 100) on model pairs (Hugging Face downloads, from model cards), over SciImpact contrastive pairs (215,928 pairs over 19 fields, built from citation counts, best paper awards and Nobel prizes, patent citations, media attention, GitHub stars and Hugging Face downloads with impact thresholds): the model sees two same-field artifacts (title and abstract, README, dataset card or model card, truncated to 1,000 words) and names the one with higher future impact in a two-option forced choice parsed by exact match; the order is balanced so chance is 50%; higher is better. Source: arxiv.org. Saturation forecast: Around 2030. 11 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-4.1 Mini57.2#789
2Qwen 2.5 14B56.2#1811
3Qwen 2.5 7B56#2929
4O4 Mini55.7#368
5nemotron-3-nano-30B-a3B54.9#1946
6Llama 3.2 3B54.8#7877
7Claude Haiku 4.554.2#583
8Qwen 3 4B54.1#2348
9Llama 3.1 8B53.5#5196
10Llama 3 8B53.4#5900
11Ministral 3 3B51.9#3504

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=sciimpact-model · How It Works · Data refreshed daily, snapshot 2026-11-21.