New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
ARC-AGI-2 benchmark
AI model leaderboard for the ARC-AGI-2 benchmark. Compare how large language models score on ARC-AGI-2, see the full ranking, and understand what this AI benchmark measures. Gemini 3.5 Flash currently leads with 72.1%. Abstraction and Reasoning Corpus v2 — visual abstract-pattern puzzles designed to resist memorisation.
Leaderboard
| # | Model | Organization | Score | Variant | Source |
|---|---|---|---|---|---|
| #1 | Gemini 3.5 Flash | Google DeepMind | 72.1% | official ↗ | |
| #2 | Claude Opus 4.5 | Anthropic | 37.6% | official ↗ | |
| #3 | GPT-5.5 Pro | OpenAI | 12% | parallel-compute | official ↗ |
| #4 | GPT-5.5 | OpenAI | 7.2% | official ↗ |
Frequently asked questions about ARC-AGI-2
What is the ARC-AGI-2 benchmark?
Abstraction and Reasoning Corpus v2 — visual abstract-pattern puzzles designed to resist memorisation.
How is the ARC-AGI-2 benchmark scored?
ARC-AGI-2 is scored using the accuracy (%) metric, where a higher score is better. The maximum achievable score is 100.000. GenAIList aggregates reported scores from model providers and papers into a single ranked leaderboard.
Which AI model scores highest on ARC-AGI-2?
As of the latest reported scores on GenAIList, Gemini 3.5 Flash achieves the highest result on ARC-AGI-2 with a score of 72.1%.
Is a higher ARC-AGI-2 score better?
Yes. On ARC-AGI-2 a higher score indicates better performance, so models near the top of the leaderboard are the strongest.