Google Unveils Gemini 3 AI Model and Antigravity Coding Interface

Google released Gemini 3, its newest large language model, making it instantly available through the Gemini app and AI search. The launch follows recent releases from competitors and introduces a research‑focused variant called Gemini 3 Deepthink for Ultra subscribers. Gemini 3 achieved record scores on the Humanity’s Last Exam and LMArena benchmarks, highlighting a notable jump in reasoning ability. Alongside the model, Google introduced Antigravity, a Gemini‑powered coding environment that blends chat, terminal, and browser windows to assist developers. The rollout aims to expand the model’s reach among millions of users and developers worldwide.

Launch Overview

Google announced the release of Gemini 3, the latest iteration of its foundation models, on Tuesday. The model is now directly accessible via the Gemini app and integrated into Google’s AI search interface. This release arrives a short period after recent announcements from OpenAI and Anthropic, underscoring the rapid pace of frontier model development.

Model Variants and Availability

A research‑intensive version named Gemini 3 Deepthink will be offered to Google AI Ultra subscribers in the coming weeks, pending additional safety testing. The base Gemini 3 model is immediately available to the broader public.

Performance Benchmarks

Independent evaluations show Gemini 3 achieving a score of 37.4 on the Humanity’s Last Exam benchmark, surpassing the previous record of 31.64 held by GPT‑5 Pro. The model also topped the LMArena leaderboard, a human‑led benchmark that measures user satisfaction. According to a Google spokesperson, the Gemini app currently hosts more than 650 million monthly active users, and 13 million developers have incorporated the model into their workflows.

New Coding Interface: Antigravity

Google introduced Antigravity, a Gemini‑powered coding interface that merges a ChatGPT‑style prompt window, a command‑line interface, and a browser view. The tool aims to provide multi‑pane, agentic coding assistance, allowing the model to interact across editors, terminals, and browsers. DeepMind CTO Koray Kavukcuoglu said, "The agent can work with your editor, across your terminal, across your browser to make sure that it helps you build that application in the best way possible."

Strategic Context

The Gemini 3 launch follows Google’s previous release of Gemini 2.5 and comes less than a week after OpenAI rolled out GPT 5.1, while Anthropic released Sonnet 4.5 two months earlier. This competitive timing highlights Google’s intent to solidify its position in the AI market.

Future Outlook

Google plans to continue expanding Gemini 3’s capabilities and to refine the Antigravity environment based on developer feedback. The combination of record‑setting benchmark performance and a dedicated coding interface positions Gemini 3 as a pivotal tool for both general AI users and software engineers.