AgentIdeaBench - Active - Specificity: leaderboard
Specificity dimension of AgentIdeaBench's Active track, where the model gets a subfield name and a literature search tool and decides what to look up (arXiv 2609.07611v1). Mean LLM-critic score on a 1-10 scale, dropping the highest of three critics and averaging over ideas and subfields, from the dims field of reports/leaderboard_full.json; higher is better.
Metric: Specificity Critic Score (1-10). Source: moyunxiang.com. Saturation forecast: Around February 2027. 63 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5 | 7.63 | #16 |
| 2 | GPT-5.6 Sol | 7.47 | #22 |
| 3 | GPT-5.6 Terra | 7.37 | #40 |
| 4 | Claude Opus 4.7 | 7.3 | #51 |
| 5 | Claude Opus 4.8 | 7.24 | #39 |
| 6 | Claude Sonnet 5 | 7.21 | #65 |
| 7 | GPT-5.5 | 7.12 | #33 |
| 8 | GPT-5.2 | 7.11 | #115 |
| 9 | GPT-5 | 7.06 | #105 |
| 10 | GPT-5.4 | 6.99 | #83 |
| 11 | GPT-5 Mini | 6.97 | #193 |
| 12 | GPT-5.6 Luna | 6.91 | #79 |
| 13 | GLM-5.1 | 6.71 | #109 |
| 14 | GPT-5.4 Mini | 6.68 | #228 |
| 15 | Gemini 3 Flash (Preview) | 6.6 | #90 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=agentideabench-active-specificity · How It Works · Data refreshed daily, snapshot 2026-11-05.