OpenAI Unveils GPT-5.5, Powering Its Push for an AI Superapp Future

OpenAI unveiled GPT-5.5 on Thursday, presenting its newest artificial intelligence model as its "smartest and most intuitive to use model" to date. This advanced algorithm boasts enhanced capabilities across numerous domains, with OpenAI co-founder and president Greg Brockman asserting that it also propels the company closer to realizing its vision for an OpenAI "superapp."

During a press briefing, Brockman characterized the new model as a substantial leap "towards more agentic and intuitive computing."

"This model is a real step forward towards the kind of computing that we expect in the future — but it is one step, and we expect to see many in the future," Brockman stated. He further elaborated, "It’s a faster, sharper thinker for fewer tokens compared to something like 5.4. So this means that there’s just more frontier AI available for businesses and for consumers, which is part of our goal."

Brockman also highlighted the model's contribution to developing a "superapp"— envisioned as a versatile, multi-purpose program akin to a Swiss Army knife. This concept, previously discussed by Brockman and co-founder Sam Altman, involves integrating services like ChatGPT, Codex, and an AI browser into a unified platform designed to support enterprise clients. Notably, the "superapp" ambition is also a prominent topic for Altman's rival and former OpenAI colleague, Elon Musk, who has expressed intentions to transform X (formerly Twitter) into its own rendition of a superapp.

The release of GPT-5.5 follows a rapid succession of model introductions, with OpenAI having launched its previous model just last month, preceded by releases in December and November. This accelerated development pace is expected to continue, according to company staff. Jakub Pachocki, OpenAI’s chief scientist, commented, "We see pretty significant improvements in the short term, extremely significant improvements in the medium term." He added, "In fact, I would say, like, I think the last two years have been surprisingly slow."

OpenAI states that GPT-5.5 is engineered for broad applicability, spanning fundamental enterprise functions such as agentic coding and knowledge work, as well as more experimental AI applications in mathematics and scientific research. The company also presented data on Thursday illustrating the model's superior performance across various benchmarks. According to OpenAI, GPT-5.5 consistently outperforms its predecessors and competitor models like Google's Gemini 3.1 Pro and Anthropic's Claude Opus 4.5.

The enduring rivalry between OpenAI and Anthropic was a recurring theme, with GPT-5.5 providing another opportunity for comparison. During the press briefing, a reporter inquired whether GPT-5.5 would possess capabilities akin to Mythos, Anthropic's recently announced cybersecurity tool, which has faced recent controversy regarding unauthorized access. Mia Glaese, a member of OpenAI’s technical staff, affirmed that GPT-5.5 would significantly influence the company's strategy for deploying its models in digital defense. "We have a strong and long standing strategy for our approach to cyber, and we’ve refined a durable approach to rolling out models safely," Glaese explained.

Mark Chen, OpenAI's chief research officer, noted that GPT-5.5 demonstrates enhanced proficiency in computer-based tasks compared to earlier models. He also highlighted that the model "shows meaningful gains on scientific and technical research workflows," expressing the company's belief that it could substantially "help expert scientists make progress." Chen further suggested its potential to assist in drug discovery, an area that has garnered increasing industry interest in recent years.

GPT-5.5 is now broadly available as of Thursday, OpenAI confirmed. The company stated that the model is being deployed to Plus, Pro, Business, and Enterprise users within ChatGPT, while GPT-5.5 Pro is specifically targeting Pro, Business, and Enterprise user tiers.