Models · The Decoder ·
OpenAI claims GPT-5.6 Sol beats Opus 5 on ARC-AGI-3 with its latest API and two additional settings
OpenAI says GPT-5.6 Sol scored 38.3% on ARC-AGI-3 using its latest API and two extra settings, beating Opus 5. In the official test setup, it scored 7.8%; ARC Prize says the environment is provider-neutral, while OpenAI alleges an outdated API affected the comparison.