Press Space to continue
GPT-6 Astra saturates the ARC-AGI 3 benchmark
Press Space to continue
ARC-AGI 3 is such a brutal bench that I’d honestly call it malicious in ways. I never thought we would even break like 30%. GPT-6 Astra saturated it. What the fuck.
ARC Prize@arcprizeSep 3
GPT-6 Astra by @OpenAI achieves SOTA on ARC-AGI: - Astra scores 63% on ARC-AGI-3, 99% via a new provider adapter harness - It surpasses human performance on 96% of ARC-AGI-3 levels - It builds the most precise symbolic model of novel environments we've seen Our analysis:
