DEEPSYNTH: leaderboard
Deep information-synthesis benchmark for agents that must gather, browse, extract, and reason over multiple sources to produce structured answers.
Metric: F1 (imported). Source: agentdeepsynthesis.github.io. Source: benchmarklist.com. Saturation forecast: Around February 2027. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O3 Deep Research | 8.97 | #115 |
| 2 | GPT-5.2 Pro | 8.7 | #61 |
| 3 | Gemini 2.5 Pro | 6.25 | #152 |
| 4 | GPT-5.1 | 3.83 | #135 |
| 5 | GPT-4.1 | 3.46 | #248 |
| 6 | O3 | 3.29 | #124 |
| 7 | O4 Mini | 3.05 | #180 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=deepsynth · How It Works · Data refreshed daily, snapshot 2026-10-20.