LaunchAI ModelsJuly 9, 2026
OpenAI launches GPT-5.6, Sol Ultra proves math conjecture
GPT-5.6 Luna outperforms GPT-5.5 at highest reasoning at 25x lower cost. Sol Ultra proved a 50-year-old math conjecture. The model tops DeepSWE at 73% and is integrated in Microsoft 365 Copilot.
15 sources
GPT-5.6 is now the preferred model in Microsoft 365 Copilotopenai.com
Summary of METR's predeployment evaluation of GPT-5.6 Solmetr.org
RT @polynoamial: GPT-5.6 Sol Ultra produced a proof of a 50 year old math conjecture. Unlike the...x.com
Meet a mathematician solving previously unsolvable math problems with GPT-5.6youtube.com
GPT 5.6 Has 72 Possible Configurations. What's A Good Default?sebastianraschka.com
OpenAI GPT-5.6: AI Could Do Anything, Then It Met ARC-AGI-3thealgorithmicbridge.com
The new GPT-5.6 family: Luna, Terra, Solsimonwillison.net
OpenAI's newest AI model is 54% more token efficient on agentic coding, Altman tells CNBCcnbc.com
AI #176 Part 1: Doing It Livethezvi.substack.com