GEMINI

2026-10

Gemini Predictions Command Center

October 2026. Gemini's prediction markets platform received an AI-powered Command Center, delivering personalized market intelligence and portfolio signals directly inside the app.

// FULL RELEASE NOTES

2026-08

Imagen 4 and Gemini 3 Image Models Shut Down

August 17, 2026. Google shut down several Imagen 4 and Gemini 3 Image models, recommending users migrate to newer stable or preview endpoints.

2026-07

Gemini 3.5 Pro Release Postponed to July 17

July 17, 2026. Google postponed the release of its anticipated Gemini 3.5 Pro model to July 17, 2026, originally expected in June, to refine its performance in complex tasks.

2026-07

Gemini in Google Sheets Promotional Limits End

July 15, 2026. Promotional access to higher limits for Gemini in Google Sheets ended, with per-user usage limits now applying.

2026-07

Gemini 3.1 Flash-Lite Preview Discontinued

July 9, 2026. The preview version of Gemini 3.1 Flash-Lite (gemini-3.1-flash-lite-preview) was discontinued, with users advised to update to gemini-3.1-flash-lite.

2026-07

Gemini Omni Flash Broader Rollout to Consumer Products

July 2, 2026. Gemini Omni Flash, Google's multimodal model for video creation and editing, saw a broader rollout to the Gemini app, Google Flow, and YouTube Shorts.

2026-07

Developer logs support for the Interactions API

July 6, 2026. Developer logs for supported Interactions API calls are now viewable in the AI Studio dashboard.

2026-06

Veo Video Generation Models Shut Down

June 30, 2026. Google shut down several Veo video generation models (veo-2.0-generate-001, veo-3.0-generate-001, veo-3.0-fast-generate-001).

2026-06

Gemini Omni Flash in public preview

June 30, 2026. Google released `gemini-omni-flash-preview`, a high-performance multimodal model for high-speed video generation and conversational video editing, in public preview.

2026-06

Nano Banana Image Generation Models

June 30, 2026. Google introduced Nano Banana 2, Nano Banana Lite, and Nano Banana Pro, a new suite of powerful and efficient image generation and editing models.

2026-06

Gemini 3.5 Flash Gains Computer Use Capability

June 24, 2026. Google added computer use capability to its Gemini 3.5 Flash model, enhancing its ability to interact with and operate digital environments.

2026-06

Gemini in Google Sheets: Formula Error Fixing

June 22, 2026. Gemini in Google Sheets can now diagnose and fix formula errors with a single click, providing explanations and corrected formulas.

2026-06

Gemini API Key Restrictions

June 19, 2026. Google began enforcing API key restrictions for the Gemini API, requiring keys to be restricted to prevent unauthorized usage and billing risks.

2026-06

Gemini 2.5 Pro with Deep Think

June 22, 2026. Google launched Gemini 2.5 Pro with Deep Think reasoning mode, its most capable model to date, featuring a 2 million token context window and leading benchmarks in science, math, and general reasoning.

2026-06

Veo 3.1 Video Generation Model

Google introduced Veo 3.1, its latest AI video generation model, which is the most capable for realistic scene and character generation.

2026-06

Gemini Omni Flash Rolls Out to Consumer Products

June 19, 2026. Google began rolling out Gemini Omni Flash, its multimodal model for video creation and editing, to the Gemini app, Google Flow, and YouTube Shorts for AI Plus, Pro, and Ultra subscribers.

2026-06

Gemini 2.5 Pro Receives Significant Update

June 19, 2026. Google pushed a significant update to Gemini 2.5 Pro, improving reasoning depth, code generation accuracy, and instruction following.

2026-06

Gemini 2.5 Flash Becomes Default Model

June 19, 2026. Gemini 2.5 Flash became the default model powering all Gemini consumer products, including the Gemini app, Gemini in Search, and Gemini in Workspace.

2026-06

Image and Video Generation Models Deprecated

June 15, 2026. Google announced the deprecation of several Imagen 4 and Gemini 3 Image models, as well as Veo video generation models and the GMP Contextual View tool.

2026-06

Gemini 2.0 Models Shut Down

June 17, 2026. Google shut down several Gemini 2.0 models, including Flash and Flash-Lite variants, recommending users migrate to Gemini 3.5 Flash or Gemini 3.1 Flash-Lite.

2026-06

Google Home Speaker with Gemini for Home

Google launches a new smart speaker built with Gemini for Home voice assistant, offering natural, conversational interaction, multi-request handling, and memory across conversations.

2026-06

Gemini API Price Reduction and Free Tier Upgrade

June 11, 2026. Google restructured its pricing for Gemini models, reducing paid API pricing by an average of 25% and upgrading the free tier to include 100 requests per day and access to Gemini 3.0 Flash.

2026-06

Gemini Features for Small Businesses

June 10, 2026. Google rolls out new Gemini features for small businesses, including connecting Google Business Profile and dedicated "Notebooks" for business organization and analysis.

2026-06

Gemini 3.5 Live Translate

June 9, 2026. Google introduces Gemini 3.5 Live Translate, a new audio model for live speech-to-speech translation across 70+ languages, available in the Gemini Live API, Google Meet, and the Translate app.

2026-06

NotebookLM Chat Upgrade with Gemini 3.5

NotebookLM's chat tool now runs on Gemini 3.5 and Antigravity, improving accuracy and providing clearer visibility into reasoning and outputs. It also gains new agentic capabilities and over 100 curated software skills.

2026-06

Gemini 3.1 Flash Audio Models (Live Preview & TTS Preview)

Google introduces Gemini 3.1 Flash Live Preview for high-quality, low-latency audio-to-audio (A2A) and Gemini 3.1 Flash TTS Preview for powerful, low-latency speech generation.

2026-06

Gemini Spark — On-Device Model for Android and Pixel

Gemini Spark is introduced as an ultra-lightweight model optimized for on-device inference on Pixel phones and Android devices, part of the broader Gemini family in 2026.

2026-06

Gemma 4 12B — Unified Multimodal Open-Weight Model

Google introduces Gemma 4 12B, a new unified, encoder-free multimodal model in its open-weight family, designed for advanced reasoning and agentic workflows.

2026-05

Googlebook Launch

May 2026. Google launched Googlebook, a new laptop experience designed for Gemini Intelligence, featuring Magic Pointer, custom widgets, and cross-device features.

2026-05

Google Health App and Fitbit Air Launch

May 2026. Google broadened its personal wellness tools with the launch of the new Google Health app and Fitbit Air.

2026-05

Gemini 3.5 Flash + Gemini Omni — Google I/O 2026

Announced May 19, 2026 at Google I/O. Gemini 3.5 Flash surpasses 3.1 Pro on coding and agentic tasks. Gemini Omni adds native video generation from any input — image, audio, video, or text.

2026-05

Gemini 3.1 Flash-Lite GA

May 2026. Gemini 3.1 Flash-Lite, Google's most cost-efficient model optimized for low-latency, high-volume tasks, is now generally available, superseding its preview version.

2026-04

Gemma 4 — Advanced Reasoning Open Weights

Released April 2, 2026. Google's fourth open-weight model family — purpose-built for advanced reasoning and agentic workflows. Surpasses Gemma 3 on most benchmarks.

2026-02

Gemini 3.1 Pro

Preview launched February 19, 2026. Strongest Gemini model available at launch — #1 on multiple coding and reasoning leaderboards. Directly challenges GPT-5.2 and Claude Opus 4.6.

2025-12

Gemini 3 Flash — New Default Model

December 2025. Speed-optimised variant of Gemini 3 becomes the default model in the Gemini app and API. Replaces 2.5 Flash as the recommended everyday model.

2025-11

Gemini 3 Pro + Deep Think

Released November 18, 2025. Google's flagship model for Q4 2025 — replaces 2.5 Pro. Deep Think variant adds extended reasoning with a visible thought chain. Competitive with GPT-5.2 at launch.

2025-05

Project Astra — Universal AI Agent

Google ships early access to Project Astra — a persistent real-world AI agent with live camera, memory across sessions, and action across all Google apps.

2025-06

Gemini 2.5 Pro GA

#1 on coding and reasoning benchmarks at launch. Native thinking mode with configurable token budget. 1M token context window standard.

2025-04

Gemini 2.5 Flash

Preview April 17, 2025 — GA June 2025. Fastest and most cost-efficient model in the 2.5 family. Competitive with 2.5 Pro on most tasks at a fraction of the cost.

2024-12

Gemini 2.0 Launch + Flash Experimental

Full Gemini 2.0 agentic family ships with native audio output, code execution, and live streaming. 2.0 Flash Experimental is the fastest model in the family.

2024-05

Gemini 1.5 Pro — 1M Context Window

1 million token context broadly available in the API. Process entire codebases or hour-long videos in a single call. Largest public context window at launch.