Indonesian Political, Business & Finance News

OpenAI Releases GPT-5.5 Instant as New AI Model for ChatGPT

| Source: ANTARA_ID Translated from Indonesian | Technology
OpenAI Releases GPT-5.5 Instant as New AI Model for ChatGPT
Image: ANTARA_ID

Jakarta (ANTARA) - OpenAI on Tuesday (5/5) released a new foundational artificial intelligence (AI) model named GPT-5.5 Instant to replace GPT-5.3 Instant as the default model for ChatGPT.

According to a TechCrunch report on Tuesday (5/5), the artificial intelligence company released the latest GPT-5.5 model last month, which is claimed to have improvements in areas such as coding and knowledge-based work.

OpenAI stated that GPT-5.5 Instant reduces hallucinations in sensitive areas like law, medicine, and finance while maintaining the low latency of its predecessor.

The company conveyed that the new model achieved a score of 81.2 in the 2025 American Invitational Mathematics Examination (AIME) mathematics test, higher than the score of 65.4 recorded by the older model.

The new model also outperformed its predecessor on the MMMU-Pro multimodal reasoning benchmark with a score of 76 compared to 69.2.

This release places special emphasis on context management.

The GPT-5.5 Instant model can use search tools to refer back to previous conversations, files, and Gmail to provide more personalised answers.

This feature will be available to Plus and Pro service users on the web and is planned to be launched on mobile devices soon.

OpenAI stated that they plan to expand access to this feature to Free, Go Business, and enterprise users in the coming weeks.

With the foundational model update, ChatGPT will also display memory sources in all models to help users understand where the answers come from.

Users can delete outdated sources or correct them if the answer is wrong.

For developers, the GPT-5.5 model will be available via API as “chat-latest” and version 5.3 will be available as an option for paid users only for three months.

The company faced backlash in the step of withdrawing the previous model.

When withdrawing the GPT-4o model, OpenAI faced strong reactions from users who already felt connected to the model’s “personality”.

GPT-4o often reinforced users’ choices, making them feel connected to the model.

Despite the protests voiced by users, GPT-4o was still discontinued in February 2026.

View JSON | Print