Introduction
Gemini has changed a lot since May 2026. Google's release notes say Gemini 4 Argon started rolling out to Google AI Pro and Ultra subscribers on 13 October 2026. Ultra now starts at $99.99 a month instead of $249.99, and Google's subscription page says the Pro plan comes with 5 TB of storage, up from 2 TB. This review covers what Gemini does today, what each plan costs according to Google's own pages, and where it falls short based on the evidence.
scored.tools rates Gemini 8.8 out of 10. We scored it against our rubric on 29 September 2026, two weeks before Argon shipped. Each criterion below is backed by a named source. We do not test tools hands-on, so this review reports what Google documents and what third parties have published.
| Criterion | Score | Evidence |
|---|---|---|
| Output quality | 9 | At the Gemini 3 launch, VentureBeat reported it ranked first on the LMArena leaderboard (Elo 1501) across text, vision, coding and web development. |
| Ease of use | 9 | The free tier at gemini.google.com needs no credit card, and the chat interface handles file uploads and Drive. |
| Pricing value | 8 | There is a free tier and paid plans from $4.99 a month. Lorphic reports the API free tier was cut by 50 to 80% in December 2025. |
| API and integration | 9 | Google's Interactions API is generally available with Python and JavaScript SDKs, REST and gRPC access, and integrations such as LiteLLM and Agno. |
| Problem fit | 9 | G2 rates it 4.4/5 and cites 900M+ monthly users, with Gemini Enterprise deployed at 2,800 companies. |
Key Features
Models and the 1M-token context window
Each plan gets different models. Google's subscription page lists Flash-Lite for Free users, Flash on AI Plus, and Pro, Flash and Gemini 4 Argon on AI Pro and Ultra. Google calls Argon its newest frontier model for complex tasks. Very little independent comparison has been published since it launched on 13 October, so treat any leaderboard claim about Argon as provisional for now.
The Pro plan FAQ lists a 1M-token context window. That is the main reason many developers and analysts pick Gemini. A large codebase, a long contract or a stack of research PDFs fits in a single conversation, so you don't have to chunk it and stitch the answers back together.
Deep Research and Search grounding
Deep Research is available on the Free plan, per the subscription page. It browses many sources and writes a cited report. Answers can also be grounded in live Google Search results. For questions about current events, prices or recent releases, that grounding is the strongest reason to pick Gemini over a chatbot that relies on a smaller search index.
Workspace, Android and the desktop
Gemini appears inside Gmail on AI Plus, and inside Gmail and Docs on AI Pro. Business editions extend it across Workspace with admin controls. On Android it works as a system-level assistant. A Windows app launched on 10 September 2026: Alt + Space opens Gemini from anywhere, and it can connect to Gmail and Drive. On 13 October Google also released Customize, which lets you tune response style and connect more than 40 apps on every tier. Google says most of those connections are US English only for now.
Images and video
All tiers get image generation and editing through Nano Banana. The video picture has changed since our May data. The Plus plan lists video generation, but Pro now gets only a limited trial of Veo 3.1 Lite, according to the FAQ. Full Veo 3.1 is listed under Ultra. If native video was your reason for buying Pro, check the current plan page before you pay.
Notebook, Skills and Antigravity
NotebookLM-style features, now labelled Gemini Notebook on the subscription page, turn uploaded sources into summaries and Audio Overviews, which are podcast-style discussions. Plus and Ultra get higher Audio Overview limits. Skills rolled out on 30 September as reusable task instructions that can be combined. They will replace Gems, and Google says existing Gems will migrate automatically. Developers on Pro and Ultra get higher limits for Jules and Antigravity, and Antigravity is now built into the macOS app for coding agents and prototyping.
API
Developers get Google AI Studio, which is free to start with limited model access, plus a paid tier with Batch API and context caching. The Interactions API covers REST, gRPC and the main SDKs. The December 2025 cut to the free API tier matters if you prototype on free quota. Plan on paid usage once you move past toy projects.
Pricing Breakdown
These are US prices from Google's subscription page. Some third-party articles still list AI Plus at $7.99 and Ultra at $249.99, which are older figures. Trust the vendor page.
| Plan | Price | Usage | Storage | Main inclusions |
|---|---|---|---|---|
| Free | $0/mo | Baseline | 15 GB | Flash-Lite, Nano Banana images, Deep Research, Gemini Live, Canvas |
| Google AI Plus | $4.99/mo | 2x Free | 400 GB | Flash model, video generation, Gemini in Gmail and Vids, 200 Flow credits |
| Google AI Pro | $19.99/mo | 4x Free | 5 TB | Gemini 4 Argon, Pro model, 1M-token context, Gemini in Gmail and Docs, 1,000 Flow credits, YouTube Premium Lite |
| Google AI Ultra | From $99.99/mo (5x Pro) or $199.99/mo (20x Pro) | 5x or 20x Pro | 20 TB+ | Veo 3.1, 10,000 or 25,000 Flow credits, Project Genie, YouTube Premium, priority access to features such as Agent Mode |
| Gemini Enterprise Business | From $21/seat/mo | Business | 25 GiB pooled per seat | Up to 300 seats, free trial |
| Gemini Enterprise Standard/Plus | From $30/seat/mo | Higher limits | 75 GiB pooled per seat | Unlimited seats, VPC-SC, customer-managed encryption keys |
Google does not publish fixed message caps. Per the subscription page, limits are compute-based and depend on prompt complexity, which features you use and how long the chat runs. They refresh every 5 hours until you hit a weekly ceiling, and you can buy AI credits to go further. Google gives no number for the Free tier. That makes it hard to know in advance whether Pro is enough for heavy daily use, which is the main argument for starting there before considering either Ultra tier.
Pros and Cons
Pros
- A 1M-token context window on AI Pro, one of the largest among major consumer chatbots.
- Live grounding in Google Search, with Deep Research available on the Free plan.
- Built into Gmail, Docs, Android, and now native Windows and macOS apps.
- AI Pro bundles 5 TB of storage and YouTube Premium Lite, which offsets much of the $19.99 if you already pay Google for storage.
- A wide developer stack: AI Studio, the Interactions API, Antigravity and Jules.
Cons
- The plans are confusing. There are four consumer plans, Ultra comes in two variants, and Gemini Enterprise is a separate family. Google's own subscription page contradicts itself on whether AI Plus includes Argon.
- Usage limits are opaque, with no published numbers for Free and only multipliers for paid plans.
- Video generation on Pro is now a limited Veo 3.1 Lite trial. Full Veo 3.1 requires Ultra.
- Our May listing recorded that Gemini's reasoning trailed Claude and GPT-5 on the hardest tasks. Argon may close that gap, but no strong independent evidence exists yet.
- Gemini's default tone is more cautious and wordy than competitors'. Customize should help, but you have to set it up.
- The API free tier was cut sharply in December 2025, per Lorphic.
Who Is It For
Gemini fits people whose work already lives in Google accounts: Gmail, Docs, Drive and Android. Gemini can read your mail and files directly, which saves a lot of copying and pasting. It also suits anyone who regularly works with very long inputs, such as a full repository, a 300-page filing or a pile of interview transcripts, because the 1M-token window on Pro handles them in one pass.
Researchers and writers who need current information will get the most from Search grounding and Deep Research. Developers already on Google Cloud get a straight path from AI Studio prototypes to production through the Interactions API and Gemini Enterprise.
Gemini is a weaker fit if your team runs on Microsoft 365, because most of Gemini's integration advantage disappears there. It is also a weaker fit if you want short, blunt answers by default. If coding agents are your main use, compare Antigravity directly with Claude and ChatGPT before paying, because Argon has only been out for days.
Verdict
Gemini earns its 8.8. Its strengths are the Google integration, the long context window and live Search grounding. Its weaknesses are the confusing plans and the vague limits.
For most readers the recommendation is Google AI Pro at $19.99 a month. It gets you Gemini 4 Argon, the 1M-token window, Gemini in Gmail and Docs, and 5 TB of storage. If you would pay for Google One storage anyway, the AI costs very little on top. Start on the Free plan first: it needs no card and includes Deep Research, so you can see whether Gemini's answer style suits you. AI Plus at $4.99 makes sense if you mostly want Gemini in Gmail and more headroom than Free.
Skip Ultra unless you keep hitting Pro's weekly ceiling or need full Veo 3.1 for real production work. Even then, the $99.99 tier is the one to try first, since the $199.99 tier only buys more usage. Developers should budget for paid API usage from the start, given the cut to the free tier. Try Gemini free, then upgrade to Pro once the Free limits get in your way.
