Ideation Arena - Specificity: leaderboard

Metric: Elo (Bradley-Terry over expert pairwise votes). Source: arxiv.org. 24 models tracked.

Top models

#ModelScore
1GPT-5.11275.3
2Claude Opus 4.51240.3
3Claude Sonnet 4.51214.7
4Kimi K2 (Thinking)1118.6
5Mistral Large 31106.3
6DeepSeek V3.21071.3
7GLM-4.71059.9
8O31059.6
9Gemini 3 Flash (Preview)1055.3
10Grok 4 Fast1049.8
11Gemini 3 Pro (Preview)1031.5
12Grok 4962.8
13Qwen 3 Max950.2
14GPT-4o836.2

Interactive version: theaggregate.ai/benchmark?slug=ideation-arena-specificity · How It Works · Data refreshed daily, snapshot 2026-09-19.