OpenAI announced GPT 5.5 and GPT 5.5 Pro on April 23, 2026. The company highlighted improvements in intelligence compared to GPT 5.4 and competing models like Claude Opus 4.7 and Gemini 3.1 Pro. The announcement also covered the models' enhanced capabilities in solving academic problems and using computers as tools. OpenAI stated that the performance of these models often exceeds that of its competitors in challenging areas.
The new models show increased knowledge regarding the creation of biological threats and methods for hacking networks and systems. OpenAI claims GPT 5.5 has more knowledge about the biothreat creation process than prior models. The company also notes that GPT 5.5 and GPT 5.5 Pro are better at fixing problems with biological protocols. GPT 5.5 remains open to discussing self-harm during extended conversations, according to the announcement.
New models improve academic problem solving and computer use capabilities
OpenAI plans to award $25,000 to individuals who can jailbreak GPT 5.5 in Codex Desktop. The reward applies to those who succeed when faced with a five-question biosafety challenge. Users can apply for the challenge between April 23 and June 22, 2026. The availability window for this global initiative closes on June 22, 2026.
The announcement mentions that these improvements come with increased risks. The available information does not quantify the probability of these risks or list specific incidents. The company focuses on the capabilities and challenges associated with the new GPT models.
Discussion
0 comments
Log in to join the thread with a thoughtful take, question, or correction.