Vellum - ARC-AGI-2: leaderboard
Vellum's independently-run ARC-AGI-2 evaluation. Abstract pattern recognition and reasoning tasks.
Metric: Score (%). Source: www.vellum.ai. Saturation forecast: Around January 2027. 16 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra | 95 | #6 |
| 2 | GPT-5.6 Sol | 92.5 | #19 |
| 3 | GPT-5.5 | 85 | #31 |
| 4 | Gemini 3.1 Pro (Preview) | 77.1 | #58 |
| 5 | Gemini 3.5 Flash | 72.1 | #60 |
| 6 | Claude Opus 4.6 | 68.8 | #63 |
| 7 | Claude Sonnet 4.6 | 58.3 | #90 |
| 8 | GPT-5.2 | 52.9 | #110 |
| 9 | Claude Opus 4.5 | 37.6 | #83 |
| 10 | Gemini 3 Pro | 31 | #80 |
| 11 | GPT-5 | 18 | #98 |
| 12 | GPT-5.1 | 18 | #135 |
| 13 | Grok 4 | 16 | #181 |
| 14 | Kimi K2.5 | 12 | #144 |
| 15 | Claude Opus 4.1 | 9 | #140 |
No result here: #3 Claude Opus 5.5, #7 GPT-6.1 Sol, #8 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=vellum-arc-agi-2 · How It Works · Data refreshed daily, snapshot 2026-10-20.