Leaderboard
ARC-AGI-1official site ↗
2 models scored · metric: accuracy. Prices are the cheapest listed across all serving providers. Cost/run and Pts/$ are priced for the chat token profile — short turns, light history, little caching — an assistant in a product.
This benchmark has not been calibrated to a token profile, so the figures use the site default. Treat the ordering as indicative rather than tuned to ARC-AGI-1. Profile: 1.5k in · 500 out · 20% cache hit · 8k context.
Scores are facts as publicly reported by labs (see report links below and on each model page) and aggregated via models.dev. ARC-AGI-1 is maintained by its own project — Model Pulse is not affiliated with or endorsed by it.
- 1.Claude Opus 55,100
- 2.GPT-5.4 Pro771
Score reports: anthropic.com · openai.com — full source URLs are linked on each model page.
Cheapest scorer: Claude Opus 5 at $5 input /M.