GPT-6 Astra scores 62.7% on ARC-AGI-3 with the standard harness and 99.9% with a new provider adapter harness; Claude Opus 5 scored 30.2%, and GPT-5.6 Sol 7.8% (Greg Kamradt/ARC Prize)

OpenAI released GPT-6 Astra, trained on its largest-ever run using more than 100,000 GPUs at the Texas Stargate facility. The model achieved 62.7% on the ARC-AGI-3 benchmark with the standard harness and 99.9% with a new provider-adapter harness, while Claude Opus 5 scored 30.2% and GPT-5.6 Sol 7.8%. Pricing is set at $10 per million input tokens and $50 per million output tokens.