Google Gemini 3 Deep Think - ARC-AGI-2: leaderboard
The "ARC-AGI-2" subtask of Google Gemini 3 Deep Think. A benchmark result reported for Google Gemini 3 Deep Think, as self-reported by the model developer.
Metric: Score (%). Source: storage.googleapis.com. Saturation forecast: Around January 2027. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Deep Think | 84.6 | #27 |
| 2 | Claude Opus 4.6 (Thinking) | 68.8 | #63 (Claude Opus 4.6) |
| 3 | GPT-5.2 (xHigh) | 52.9 | #110 (GPT-5.2) |
| 4 | Gemini 3 Pro (Preview) | 31.1 | #66 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=google-gemini-3-deep-think-arc-agi-2 · How It Works · Data refreshed daily, snapshot 2026-10-20.