Gemini Token Plan pricing boils down to one sentence: Free defaults to Gemini 3.5 Flash, Google AI Plus ($4.99/mo) roughly doubles quota, Google AI Pro ($19.99/mo) roughly quadruples it and unlocks full Gemini 3.1 Pro, while Ultra 5x ($99.99/mo) and Ultra 20x ($199.99/mo) push to ~5× and ~20× Free. This is not a per-token API — it is a subscription quota pool for the Gemini app and the Google ecosystem. All prices and quota data below were verified against the official subscription page on 2026-08-18 (source).
Gemini Token Plan Pricing Comparison at a Glance: Five Tiers, One Table
Gemini Token Plan has five tiers — Free, Google AI Plus at $4.99/mo, Google AI Pro at $19.99/mo, Google AI Ultra 5x at $99.99/mo, and Google AI Ultra 20x at $199.99/mo — and a higher price always buys a bigger usage multiplier plus more models and tools. The Free tier includes 15GB of Google One storage, Gemini 3.5 Flash by default, and varying access to Gemini 3.1 Pro; paid tiers step up from 400GB on Plus to 20TB+ on Ultra, with bundled YouTube Premium and Google Home perks in select regions (source). Here is the full picture:
| Tier | Monthly (USD) | Usage multiplier (vs Free) | Google One storage | Key unlocks |
|---|---|---|---|---|
| Free | $0 | 1× (baseline) | 15GB | Gemini 3.5 Flash default, varying 3.1 Pro access |
| Google AI Plus | $4.99 | ~2× | 400GB | Gemini Omni, Daily Brief, video generation, 200 Flow Credits |
| Google AI Pro | $19.99 | ~4× | 5TB | Full Gemini 3.1 Pro, expanded Deep Research, Jules, Antigravity |
| Google AI Ultra 5x | $99.99 | ~5× | 20TB | Deep Think, Agent Mode, Veo 3.1 priority, 10,000 Flow Credits |
| Google AI Ultra 20x | $199.99 | ~20× | 20TB | Project Genie (US), 25,000 Flow Credits, Gemini Spark |
Keep one number in mind: Plus ~2×, Pro ~4×, Ultra 5x ~5×, and Ultra 20x ~20× Free is the core variable of this entire pricing comparison. The same workload runs roughly 4× on Pro and 20× on Ultra 20x versus Free, so estimate your daily usage before judging which tier is worth it.
Tier-by-Tier Breakdown: What Plus, Pro, and Ultra Actually Differ On
Google AI Pro ($19.99/mo) is the officially "most popular" tier and the natural stopping point for most upgraders: ~4× Free quota, full Gemini 3.1 Pro, 1M-token context, and the Jules coding agent. Tier details (source):
Google AI Plus ($4.99/mo): The Cheapest Paid Tier
Plus roughly doubles Free quota for only $4.99/mo, which mostly solves the "always hitting limits" problem. Compared with Free, Plus adds Gemini Omni, Daily Brief, video generation, more Gemini 3.1 Pro and Deep Research access, 200 Flow Credits, 400GB of storage, and early Gemini in Chrome access. It fits daily Gemini users who are not yet at research intensity.
Google AI Pro ($19.99/mo): The Capability Divide
Pro is the biggest capability jump: ~4× Free quota, and only from this tier do you get full Gemini 3.1 Pro, 1M-token context (~1,500-page uploads), expanded Deep Research, and the Jules async coding agent. Flow Credits rise from 200 to 1,000, NotebookLM limits are 5× higher, storage hits 5TB, and you get YouTube Premium Lite (select regions) plus $10/mo Google Cloud credits. Knowledge workers and research users mostly stop here.
Google AI Ultra 5x / 20x ($99.99 / $199.99/mo): Only for Extreme Use
The Ultra tiers are about "never hitting the wall": Ultra 5x runs ~5× Free usage and Ultra 20x ~20×, with priority access to Deep Think, Agent Mode, and Veo 3.1, plus Project Genie on US Ultra 20x. Flow Credits are 10,000 and 25,000 respectively, storage starts at 20TB, and Google Home Premium plus YouTube Premium individual are bundled in. Overkill for most, but closer to a "no wall" experience for all-day research, multi-tool parallelism, and heavy creation.
Quota, Usage and Rate Limits: 5-Hour Refresh plus Weekly Caps
Gemini Token Plan quota is not a fixed "N messages per month" — it is a compute-based limit that refreshes every 5 hours and accumulates toward weekly caps after each window. Consumption scales with prompt complexity, the features used (Deep Research, video generation, and so on), and chat length, so the same conversation burns a very different share of quota on Free versus Ultra 20x. When you hit a cap, either wait for the next refresh or buy AI credits for overage (source).
The multiplier decides how much you can actually use: Plus ~2×, Pro ~4×, Ultra 5x ~5×, Ultra 20x ~20× Free. In other words, whatever a Free 5-hour window lets you run, Pro runs roughly 4× as much. If you routinely burn through Free quota, estimate the tier from the multiplier rather than staring at the monthly price.
Top-up, Deals and Official Entry: Where to Subscribe, Whether Annual Billing Exists
The official entry is gemini.google/subscriptions/, billed monthly at signup; no annual or first-purchase discounts are listed in the data, and the real deals live in each tier's bundled perks. Plus brings 400GB of storage, Pro brings 5TB plus YouTube Premium Lite (select regions), and Ultra brings YouTube Premium individual — bundles vary by region (source). Overage is handled by buying AI credits inside the subscription system, which is separate from Gemini API pay-as-you-go.
Subscription vs API: Gemini Token Plan and Gemini API Are Two Billing Systems
gemini.google subscription quota never becomes API key balance — Gemini Token Plan is a consumer subscription, while Gemini API bills per token on ai.google.dev, and the two never offset each other. The subscription buys a quota pool inside Gemini App, Gmail, Docs, and Jules; self-built apps, agents, and AI Studio production calls need a separate Paid tier (source). Side by side:
| Dimension | Gemini Token Plan subscription | Gemini API pay-as-you-go |
|---|---|---|
| Billing model | Monthly subscription, quota pool + 5-hour refresh/weekly caps | Per-token billing, pay for what you use |
| Official entry | gemini.google/subscriptions/ | ai.google.dev |
| Covers | Gemini App, Gmail/Docs/Sheets, Jules, Deep Research, NotebookLM | Open API, Google AI Studio, Vertex AI |
| Free quota | Free tier: 15GB storage + basic features | AI Studio Free tier on select models (rate-limited) |
| Best for | Consumers using Gemini inside Google products | Self-built apps, agents, production integration |
For reference, API-side pricing: Gemini 3.6 Flash at $1.50 input / $7.50 output per 1M tokens, Gemini 3.5 Flash at $9 output, Gemini 3.1 Pro at $2 input / $12 output (≤200K), and the cheapest Gemini 3.1 Flash-Lite at $0.25 input (source). If you want to do the per-token math or build production integrations, check the Gemini API comparison page directly.
Is It Worth It: Recommendations by Use Case
Bottom line: heavy Google-ecosystem Gemini users are best served by Pro ($19.99/mo); light users are fine with Plus ($4.99/mo); only all-day research, creation, and coding workflows need Ultra. Recommendations by scenario (source):
| Your situation | Recommendation |
|---|---|
| Occasional use, just trying it out | Free tier, Gemini 3.5 Flash by default |
| Daily but light use, always hitting Free caps | Google AI Plus $4.99/mo (~2× quota) |
| Heavy research + Deep Research + Jules coding | Google AI Pro $19.99/mo (~4× + full 3.1 Pro) |
| All-day intensity, want Deep Think and early features | Google AI Ultra 5x $99.99/mo |
| Extreme use, multi-tool parallelism, need Project Genie | Google AI Ultra 20x $199.99/mo |
| Building apps / production integration | Gemini API pay-as-you-go on ai.google.dev |
Three final reminders: quota is computed dynamically on a 5-hour refresh plus weekly caps, not a fixed monthly count; bundled perks like YouTube Premium and Google Home vary by region, so check your local page before paying; and subscription quota cannot be spent as API key balance — the two systems stay separate. For a broader look at subscription AI assistants, see the ChatGPT and Claude comparison pages. Whether Gemini Token Plan is worth it comes down to how much quota you burn per day: stay on Plus if you live inside 2×, move to Pro if you exceed 4× and want full Gemini 3.1 Pro, and treat Ultra as mostly overkill otherwise.