AgentIdeaBench - Static: leaderboard
Research-hypothesis ideation in AgentIdeaBench's Static track, where the model is handed the references for a subfield (arXiv 2609.07611v1). Weighted LLM-critic total on a 1-10 scale (originality 2, impact 1.5, feasibility 1, clarity 0.5, specificity 0.5; highest of three critics dropped), from reports/leaderboard_full.json; higher is better.
Metric: Weighted Hypothesis Score (1-10). Source: moyunxiang.com. Saturation forecast: Around September 2028. 63 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Sol | 6.9 | #22 |
| 2 | Claude Opus 5 | 6.89 | #16 |
| 3 | GPT-5.6 Terra | 6.86 | #40 |
| 4 | GPT-5.5 | 6.7 | #33 |
| 5 | GPT-5.6 Luna | 6.7 | #79 |
| 6 | GPT-5 | 6.55 | #105 |
| 7 | GPT-5.4 | 6.18 | #83 |
| 8 | GPT-5.2 | 6.16 | #115 |
| 9 | GPT-5.4 Mini | 6.06 | #228 |
| 10 | Claude Sonnet 5 | 6.05 | #65 |
| 11 | GPT-5 Mini | 5.88 | #193 |
| 12 | Gemini 3.5 Flash | 5.87 | #66 |
| 13 | Claude Opus 4.7 | 5.83 | #51 |
| 14 | Claude Opus 4.6 | 5.8 | #67 |
| 15 | Claude Opus 4.8 | 5.78 | #39 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=agentideabench-static · How It Works · Data refreshed daily, snapshot 2026-11-05.