How enabling two settings tripled our scores on the ARC-AGI-3 benchmark
The article highlights that enabling specific API settings, such as "retained reasoning" and "compaction," in the GPT-5.6 Sol model significantly improved its performance on the ARC-AGI-3 benchmark, increasing scores from 7.8% to 38.3% and reducing output tokens by 6x. These settings allowed the model to better retain reasoning and manage memory, addressing limitations in previous versions. The study emphasizes that evaluating AI models in isolation is insufficient and that factors like API settings and harness design play a critical role in determining performance.
More from Openai
How RingCentral builds AI-native work from engineering to ops
See how RingCentral uses ChatGPT Work and Codex to accelerate AI product development and centralize operational intelligence across engineering and operations.
Daybreak models are now available on AWS
OpenAI and AWS are making Daybreak cybersecurity capabilities available through Amazon Bedrock to support enterprise security workflows.
Testing ads in ChatGPT
OpenAI begins testing ads in ChatGPT to support free access, with clear labeling, answer independence, strong privacy protections, and user control.
