Two API settings tripled scores on the ARC-AGI-3 benchmark by GPT-5.6, boosting performance and efficiency through retained reasoning and enabled compaction.
Read the original at openai.com→How two API settings improved GPT-5.6 performance on ARC-AGI-3, boosting scores and efficiency by retaining reasoning and enabling compaction.
Original headline: "How enabling two settings tripled our scores on the ARC-AGI-3 benchmark"
Coverage timeline
- Jul 29, 15:00 UTC OpenAI lead source How enabling two settings tripled our scores on the ARC-AGI-3 benchmark