Gemini is Google's flagship AI, built by Google DeepMind and deployed across nearly every Google surface — the Gemini app, Search, Workspace, Android, Pixel, Chrome. It's the product with the largest distribution footprint in consumer AI by a wide margin, and it's the strongest alternative to ChatGPT for anyone who lives inside the Google ecosystem.
This page is the living reference to Gemini — the model family, pricing, what's new, how it compares, and every story we publish tagged Gemini. We update it whenever Google ships a meaningful change.
As of April 2026, Gemini 3.1 Pro is the frontier default in the Gemini app for AI Premium subscribers. Gemini Ultra, the flagship, is reserved for the $249/month AI Ultra tier and for Vertex AI enterprise deployments.
We rated Gemini 4.5 / 5. The right AI chatbot if you live in Google Workspace. Gemini 2.5 Pro is competitive with GPT-5 and Claude on most tasks, and the Gmail/Docs/Drive integration is genuinely useful. Free tier is the most generous in the category.
Read the full Gemini review →Latest news about Gemini
Every story we've published tagged Gemini. Refreshed automatically every time the newsroom publishes.
- Aug 20, 2026Grok leaks user chats when prompt injections arrive encryptedAdversa researchers bypassed xAI's guardrails with AES-256-GCM ciphertext; xAI was told in June and the flaw still works.
- Aug 18, 2026Firefox's Smart Window adds live web search and cited AI answersMozilla is betting an opt-in, model-agnostic AI browser can beat Google's Chrome integration on privacy and user choice.
- Aug 14, 2026Google makes visible watermark on AI images, video, and audio optionalNano Banana, Omni, and Lyria outputs can now ship without the visible tag; invisible SynthID and C2PA metadata stay on.
- Aug 13, 2026Google ships Gemini 3.7 Flash at half the price of 3.6, three weeks laterThe new workhorse model posts 43.6% on FrontierCode 1.1 and 65.3% on DeepSWE v1.1, with input tokens at $0.75 per million.
- Aug 12, 2026Google unveils Pixel 11, Pixel Tag, and Gemini features at Made by Google '26The Pixel 11 starts at $899, a $100 hike; Google finally answers Apple's AirTag with a $29 Pixel Tag tied to Find Hub.
- Aug 12, 2026Google's Pixel Watch 5 leans on AI and health software, adds $50 to the priceThe $399 watch keeps the hardware nearly static and bets on offline Gemini, insulin resistance trends, and breathing emergency detection.
Timeline of major releases
- Dec 6, 2023Gemini 1.0 announcedGoogle DeepMind unifies Bard and the broader Gemini project. Initial launch includes Ultra, Pro, and Nano sizes, with Ultra positioned against GPT-4.
- Feb 8, 2024Bard renamed to GeminiConsumer surface gets the Gemini brand. Gemini Advanced subscription launches at $19.99/month with access to Gemini Ultra.
- May 14, 2024Gemini 1.5 Pro and the 1M context windowContext leap to 1 million tokens makes Gemini the default for long document and video analysis. I/O 2024 emphasizes multimodal — text, audio, and video all handled natively.
- Feb 20, 2025Gemini 2.0 with agent capabilitiesProject Astra demo becomes a shipped product. Gemini gets real-time multimodal understanding and agent-style tool use.
- Dec 8, 2025Gemini 3 launchesGemini 3 Pro and Ultra surpass prior state of the art on multimodal reasoning benchmarks. Full Google Workspace integration at default tier.
- Apr 15, 2026Gemini 3.1 Pro shipsImproved long-context retrieval and new GPQA Diamond benchmark record. Deeper Workspace integration — Gmail, Docs, Meet — in Gemini Business tier.
How to access Gemini
| Plan | Price | Includes |
|---|---|---|
| Free | $0 | Gemini 2.5 Flash in the Gemini app, limited Pro access, integration with Google Search. |
| AI Premium (Google One) | $19.99/mo | Gemini 3 Pro in the app, Gemini in Gmail/Docs/Sheets, 2TB Google One storage. |
| AI Ultra | $249/mo | Gemini Ultra with highest caps, Veo video generation priority, priority research tools. |
| Gemini Business (Workspace) | $20/user/mo (add-on) | Gemini across Workspace apps for organizations, admin controls, data residency. |
| Enterprise / Vertex AI | Usage-based | API access via Google Cloud Vertex, SOC 2, HIPAA-eligible surfaces, custom models. |
Gemini vs. ChatGPT vs. Claude
| Feature | Gemini | ChatGPT | Claude |
|---|---|---|---|
| Context window (paid) | 1M+ (2M in Advanced) | 128K | 200K (1M in API) |
| Native video understanding | Yes (best in class) | Limited | Images only |
| Productivity suite integration | Google Workspace (native) | Microsoft 365 via Copilot (separate) | Via integrations only |
| Search grounding | Yes (Google) | Yes (Bing) | Via API tools |
| Image generation | Imagen | DALL-E / GPT Image | No |
| Video generation | Veo 3 (included in paid tiers) | Sora (Pro tier) | No |
The Gemini model family
Google ships three sizes: Flash (fastest, cheapest), Pro (default), and Ultra (frontier). A fourth — Nano — runs on-device on Pixel phones for offline and low-latency uses. The Gemini app and Workspace integration pick the right size based on tier and task.
Gemini's native multimodality is the clearest technical differentiator. Unlike OpenAI and Anthropic, which retrofitted image and audio onto text-first models, Gemini was trained multimodal from the start. In practice this shows up most strongly in long video understanding — Gemini can ingest an hour of video and reason over it coherently, which no competitor matches.
How Gemini fits into Google Workspace
The most interesting Gemini story in 2026 isn't the standalone chatbot — it's the integration across Gmail, Docs, Sheets, Meet, and Drive. Paid Workspace tiers get a sidebar AI that has native context of your email, your documents, and your calendar. The productivity case is strong: summarize a thread, draft a reply, pull numbers from a spreadsheet, generate meeting notes from a recording. See our AI tools comparison for how this stacks up against Microsoft Copilot.
Pricing and plans
Google sells Gemini three ways: as a consumer subscription (AI Premium and Ultra, via Google One), as a Workspace add-on for organizations, and as a metered API via Vertex AI for developers and enterprise. The consumer pricing mirrors OpenAI and Anthropic at $20/month, and the Ultra tier at $249 is priced to match OpenAI's Pro.
The developer economics are interesting: Gemini Flash, as of 2026, is the cheapest frontier-class model on the market, routinely beating GPT-5 mini and Claude Haiku on price per useful token. For high-volume workloads where latency and cost matter more than absolute peak capability, Flash is increasingly the default.
Where Gemini wins — and where it doesn't
Gemini is the right choice if you live in Google Workspace, if your work involves long video or long document reasoning, or if you need the absolute longest context window available in a shipping consumer AI (1M+ tokens, 2M in Advanced). It's also the default AI experience on Android, and increasingly on Google Search, so for most people the question isn't "should I use Gemini" — it's "which AI do I use beyond Gemini."
It trails Claude and ChatGPT on coding (Claude Code is the reference developer product, Codex and GPT-5 agents are closing the gap). It trails ChatGPT on real-time voice for most use cases (Gemini Live is competitive but Advanced Voice still feels more natural). And Google's pace of product changes — model renames, subscription restructurings, surface consolidations — is higher than its competitors', which can be frustrating for teams trying to standardize.
Our coverage
Everything tagged Gemini appears in Latest news above. For Google's broader AI strategy, see our reporting under the AI news and AI business beats. For head-to-head alternatives, see ChatGPT and Claude. For the underlying agent capabilities, see our AI agents guide.
