AI provider / Google
A provider often linked to broad AI platform reach, multimodal research, consumer products and both proprietary and open model lines.
Google pushes AI across research, cloud, search and global consumer products, and runs one of the broadest model strategies in the market. Gemini is the flagship multimodal line — from the million-token context of 1.5 Pro to the thinking-by-default Gemini 2.5 and the day-one ubiquity of Gemini 3.
In parallel, the lighter Gemma family carries that same research into open weights, giving developers an adaptable, single-GPU starting point. Hosted frontier on one side, open and portable on the other — that dual track defines Google’s position.
Model catalog
Every release in this provider’s model line.
Each row is a distinct model release — its date, its type and the traits it is known for. Open a source to dig into any of them.
| Model | Released | Type | Key traits | Source |
|---|---|---|---|---|
| Gemini 4 Argon | Sep 2026 | Flagship |
| |
| Gemini 3.8 Flash | Sep 2026 | Version update |
| |
| Gemini 3.5 Flash | May 2026 | Agentic |
| |
| Gemma 4 | Apr 2026 | Open release |
| |
| Gemini 3 Flash | Dec 2025 | Version update |
| |
| Gemini 3 Pro | Nov 2025 | Flagship |
| |
| Gemini 2.5 Pro | Mar 2025 | Reasoning |
| |
| Gemma 3 | Mar 2025 | Open release |
| |
| Gemini 2.0 Flash | Dec 2024 | Agentic |
| |
| Gemma | Feb 2024 | Open release |
| |
| Gemini 1.5 Pro | Feb 2024 | Version update |
| |
| Gemini 1.0 | Dec 2023 | Debut |
|
Timeline
A practical timeline of model releases and company milestones.
This page tracks the moments that matter most: major model releases, version jumps and company moves that changed how this provider shows up in the AI market.
Gemini 4 Argon opens a new frontier era, trained for cyber defense.
With a 1M-token context window, Argon targets real-world coding, enterprise knowledge work and autonomous vulnerability patching, rolling out first to vetted cyber defenders before opening to Pro and Ultra subscribers on October 13.
Gemini 3.8 Flash is Google’s third Flash release in six weeks.
Billed as the best reasoning and coding model yet at 3.7 Flash’s price, it ships alongside a Flash Cyber variant gated behind the Fairwind trusted-access program.
Gemini 3.5 Flash bets Google’s next wave on agents, not chatbots.
Announced at I/O 2026 with built-in computer use and 4x the output speed of comparable frontier models, Flash becomes Google’s lead vehicle for long-horizon, tool-using work.
Gemma 4 brings Gemini 3-era research to open weights.
Released under Apache 2.0, the fourth generation keeps Google competitive in the open ecosystem against Llama and Qwen.
Gemini 3 Flash becomes the new default in the Gemini app.
The fast tier inherits most of Pro’s capability, with the February Deep Think upgrade later targeting science and engineering work.
Gemini 3 Pro ships everywhere on day one.
Launched simultaneously across Search, the Gemini app, AI Studio and Vertex, with Deep Think following for the hardest problems.
Gemini 2.5 Pro makes “thinking” the default flagship behavior.
Google’s first explicitly designated thinking model tops leaderboards and folds the 2.0 experiments into one reasoning line.
Gemma 3 adds multimodality and 128K context to the open line.
Four sizes from 1B to 27B, single-GPU friendly and agent-ready, after Gemma 2 had matured the family in 2024.
Gemini 2.0 Flash opens the agentic era for Google.
Built for tool use and native audio-image output, it arrives alongside Project Astra and Mariner agent prototypes.
Gemma opens a lightweight branch built on Gemini research.
The 2B and 7B open models give Google a stake in the open-weight ecosystem dominated by Llama and Mistral.
Gemini 1.5 Pro debuts the million-token context window.
MoE architecture plus breakthrough long-context recall make whole-codebase and hour-long-video prompts practical for the first time.
Gemini 1.0 launches in Ultra, Pro and Nano sizes.
Google’s answer to GPT-4 is natively multimodal from day one and immediately spans cloud, app and on-device tiers.