Provider of the Gemini model family and Vertex AI.
At a glance
Google AI history
Google has been at the forefront of AI research since the early 2010s. Its 2017 paper "Attention Is All You Need" introduced the Transformer architecture, which underpins every modern large language model — including those of its competitors.
DeepMind acquisition
In 2014, Google acquired London-based DeepMind for £400 million. DeepMind contributed AlphaGo (2016), AlphaFold (2020 — Nobel-prize-winning protein structure prediction), and major work on reinforcement learning. In 2023, Google merged its Brain and DeepMind teams into Google DeepMind under Demis Hassabis.
The Gemini family
After the launch of ChatGPT, Google accelerated its consumer-AI strategy:
- Bard (Mar 2023) — initial chatbot, later renamed
- Gemini 1 (Dec 2023) — first natively multimodal model family
- Gemini 1.5 (Feb 2024) — 1M-token context window
- Gemini 2.0/2.5 (2024–2025) — current frontier models with strong multimodal performance
- Imagen, Veo — image and video generation
Cloud & APIs
Google offers AI through:
- Vertex AI (Google Cloud Platform) — enterprise endpoints
- AI Studio — free developer playground
- Gemini API — direct REST API
- Gemini app — consumer-facing chatbot
Ad business
Google's AI investments are funded by Alphabet's advertising business, which generates over US$200 billion annually.
Latest activity
Full history →- Pricing
Gemini 2 Flash price cut 50%
Google halves Gemini 2 Flash pricing to $0.10 input / $0.40 output per 1M tokens.
Gemini 2 Flash - Release
Gemini 2 Pro and Flash released
Google releases Gemini 2 Pro with 2M context and Gemini 2 Flash for high-throughput use.
Gemini 2 Pro
Releases timeline
Showing 9 most recent- Gemini 2.5 Flash
Fast and efficient multimodal model.
- Gemini 2.5 Flash Lite
Ultra-fast lightweight variant.
- Gemini 2.5 Pro
Google Gemini 2.5 Pro — state-of-the-art thinking model.
- Gemini 2 Flash
Low-latency, low-cost multimodal model with 1M context.
- Gemini 2 Pro
Google's flagship multimodal model with very long context.
- Gemma 2 27B
Open-weights 27B model from Google with state-of-the-art performance at its size.
- Gemma 2 9B
Open-weights 9B model from Google, competitive with much larger models.
- Gemini 1.5 Flash
Fast multimodal model from the Gemini 1.5 generation.
- Gemini 1.5 Pro
Previous Google flagship with 1M context window, superseded by Gemini 2.
Active models
- Gemini 2 FlashActive
Low-latency, low-cost multimodal model with 1M context.
1.0M ctx - Gemini 2 ProActive
Google's flagship multimodal model with very long context.
2M ctx - Gemini 2.5 FlashActive
Fast and efficient multimodal model.
1.0M ctx - Gemini 2.5 Flash LiteActive
Ultra-fast lightweight variant.
1.0M ctx - Gemini 2.5 ProActive
Google Gemini 2.5 Pro — state-of-the-art thinking model.
1.0M ctx - Gemma 2 27BActive
Open-weights 27B model from Google with state-of-the-art performance at its size.
8K ctx - Gemma 2 9BActive
Open-weights 9B model from Google, competitive with much larger models.
8K ctx
Older & deprecated (2)
Community ratings
Rate Google
Sign in to rate and review.
Comments
Sign in to leave a comment.