OnenAI launched the GPT 5.5 model last month, claiming improvements in areas such as coding and professional tasks. Now, the AI firm is back with another AI model, GPT-5.5 Instant on Tuesday. This new model is set to replace the default GPT-5.3 Instant, which was launched just two months back on 3rd March, 2026.
What the new update brings to users
The company claims improvements in areas such as medicine, law, and finance, stating, “GPT-5.5 Instant produced 52.5% fewer hallucinated claims than GPT-5.3 Instant on high-stakes prompts covering areas like medicine, law, and finance.” It also said inaccurate responses were reduced by 37.3% in challenging, user-flagged conversations.
The new model also shows clear gains in benchmarks. It scored 81.2 on the AIME 2025 math test, up from 65.4 on the older version. It also did better on the MMMU-Pro multimodal reasoning benchmark, reaching 76 compared to 69.2 earlier.
With this update, ChatGPT will now show where its responses are coming from by highlighting memory sources across all models. This can give users more clarity and control — as they now can review, delete outdated information, or correct anything that seems inaccurate. Importantly, OpenAI says these memory sources remain private, so if you share a chat with someone, they won’t be able to see them.
A key highlight of the update is improved context handling. GPT-5.5 Instant can now use its search capabilities to reference past conversations, files, and even Gmail data to deliver more personalised responses.
Chat GPT 5.5 Instant availability
The GPT 5.5 Instant is currently rolling out to Plus and Pro users on the web, with mobile support expected soon. OpenAI has also indicated plans to extend access to Free, Go, Business, and enterprise users in the coming weeks.
For developers, GPT-5.5 is available via API under the “chat-latest” model, while GPT-5.3 — before it gets retired — will continue to be offered as an option for paid users for a limited period of three months.
Alongside these updates, GPT-5.5 is designed to handle a wider range of everyday tasks beyond just coding. It can assist with research, data analysis, and document creation, while managing multi-step workflows more efficiently.
