LLM Stats (ARC-AGI-3): leaderboard
LLM Stats mirror of ARC Prize’s interactive-reasoning benchmark: an agent is dropped into a novel game environment with no instructions, no stated goal and no rules, and has to infer all three by playing. Six environments of eight to ten levels each, scored on levels completed with total actions as the tiebreaker — humans clear every one of them.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Around December 2027. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5 | 30.2 | #12 |
| 2 | GPT-5.6 Sol | 7.78 | #19 |
| 3 | GPT-5.6 Terra | 0.8 | #41 |
| 4 | GPT-5.6 Luna | 0.18 | #68 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-arc-agi-3 · How It Works · Data refreshed daily, snapshot 2026-10-20.