reasoning benchmark
ARC-AGI Leaderboard
ARC-AGI is a benchmark for general intelligence and abstract reasoning using visual grid-based transformation tasks.
Updated Sep 5, 2026
Models9
Model coverage9
MetricScore
EvidenceB
ARC-AGI Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelOP | Score98.50% | Percentile100.00% | Participants9 | EvidenceC | Evaluated |
| Rank02 | ModelOP | Score95.00% | Percentile87.50% | Participants9 | EvidenceC | Evaluated |
| Rank03 | ModelOP | Score93.70% | Percentile75.00% | Participants9 | EvidenceC | Evaluated |
| Rank04 | ModelOP | Score90.50% | Percentile62.50% | Participants9 | EvidenceC | Evaluated |
| Rank05 | ModelOP | Score88.00% | Percentile50.00% | Participants9 | EvidenceC | Evaluated |
| Rank06 | ModelOP | Score86.20% | Percentile37.50% | Participants9 | EvidenceC | Evaluated |
| Rank07 | ModelTM | Score84.00% | Percentile25.00% | Participants9 | EvidenceC | Evaluated |
| Rank08 | ModelME | Score50.30% | Percentile12.50% | Participants9 | EvidenceC | Evaluated |
| Rank09 | ModelAC | Score41.80% | Percentile0.00% | Participants9 | EvidenceC | Evaluated |
ARC-AGI Highlights
The leading models and scores on this benchmark.
ARC-AGI Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for ARC-AGI
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is ARC-AGI?
What ARC-AGI measures and how its scores work.
The Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI) consists of visual grid-based tasks with 2-5 demonstration pairs and novel test inputs that require rules to be inferred and applied.
It measures general intelligence, abstract reasoning, human-like general fluid intelligence, and skill-acquisition efficiency with minimal prior knowledge, reported as a Score ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- ARC-AGI
- Modality
- image
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about ARC-AGI.
Which model scores highest on ARC-AGI?
GPT-6 Astra is currently ranked first with 98.50%.
What are the top three models on ARC-AGI?
The current leaders are GPT-6 Astra (98.50%), GPT-5.5 (95.00%), and GPT-5.4 (93.70%).
Which ARC-AGI model has the lowest official input price?
Qwen3-235B-A22B-Instruct-2507 has the lowest matched official input price at $0.70 input / $2.8 output per 1M tokens.
Which models are fastest among ARC-AGI results?
The fastest matched records are GPT-5.5 (134.94 tok/s via OpenAI), GPT-5.2 (100.00 tok/s via OpenAI), and GPT-5.4 (50.00 tok/s via OpenAI).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.