Home / Guides / How Much Does Gemini Cost
How Much Does Gemini Cost? (2026 Pricing Breakdown)
Gemini is unusual: the CLI used to be free for most developers, but since June 18, 2026 it no longer serves free individual accounts, and the price goes in two very different directions, a flat Code Assist subscription or pay-per-token API billing. Here's exactly what each one costs in 2026 and how to keep an eye on your real spend.
Last updated 7 November 2026 (Gemini app and API prices checked 16 October 2026) · By Soren Starck
The Short Answer
For individual developers, the free Gemini CLI tier is gone. Since June 18, 2026, Gemini CLI no longer serves free Gemini Code Assist for individuals, or Google AI Pro and Ultra subscribers (Google points them to its new Antigravity CLI). To keep using the Gemini CLI, there are two paths:
- Gemini Code Assist (paid): a flat monthly subscription, Standard ($19/user/mo with an annual commitment, $22.80 without) or Enterprise ($45/user/mo annual, $54 without), with higher quotas, an SLA, and admin controls.
- Gemini API: pay-per-token billing through Google AI Studio / Vertex AI. You pay only for what you send and receive, but cost scales directly with usage.
That is the entire Gemini pricing story in one paragraph. The rest of this guide breaks down each tier with real numbers and worked examples so you can predict your bill.
Gemini Pricing at a Glance
| Tier | Price | What You Get | Best For |
|---|---|---|---|
| Code Assist (Free) | $0 | No longer serves Gemini CLI or the IDE extensions (since June 18, 2026) | No longer a CLI option |
| Code Assist Standard | $19/user/mo* | Higher quotas, SLA, admin controls | Google Cloud teams |
| Code Assist Enterprise | $45/user/mo* | Highest quotas, code customization, security review | Larger orgs |
| Gemini API | Per token | Pay-as-you-go, no flat fee, Flash & Pro tiers | Apps & heavy automation |
*With an annual commitment. Without one, Standard is $22.80/user/mo and Enterprise $54/user/mo. Prices shift as Google rebalances tiers, always cross-check the official Gemini Code Assist pricing page.
Gemini App Plans: Google AI Plus, Pro and Ultra
If you mean the Gemini app rather than the developer tools, Google sells it as Google AI plans. US prices on gemini.google/subscriptions, checked 16 October 2026:
| Plan | Price | Gemini models |
|---|---|---|
| Free | $0 | Flash-Lite, with a Google account |
| Google AI Plus | $4.99/mo | Flash |
| Google AI Pro | $19.99/mo | Gemini 4 Argon, Pro and Flash |
| Google AI Ultra | $99.99 or $199.99/mo | Gemini 4 Argon; 5x or 20x Pro's usage limits |
None of these plans buys Gemini CLI access any more (see below); for coding agents, Google points AI Pro and Ultra subscribers to Antigravity, whose limits and credits are in Antigravity pricing.
The Free Tier (Gone From the CLI Since June 2026)
The free Gemini CLI tier was the most generous in its class: Gemini Flash and Gemini Pro models with no credit card, and a daily request budget most solo developers never hit. That ended on June 18, 2026, when Gemini CLI and the Gemini Code Assist IDE extensions stopped serving requests for free Gemini Code Assist for individuals and for Google AI Pro and Ultra subscribers. Access through a Code Assist Standard or Enterprise license, or a paid Gemini API key, is unchanged. What each of those routes allows a day is in Gemini CLI usage limits.
The catch is that quotas meter requests, not dollars, and there is no native warning before you run out. When you exceed your requests-per-minute or requests-per-day limit you get a 429 Resource Exhausted error, and on the Gemini API the daily quota resets at midnight Pacific time, not your local timezone. If you want the full breakdown of those caps, see our Gemini rate limits guide.
Gemini Code Assist (Paid Subscriptions)
With the free individual tier gone from the CLI, a Code Assist subscription is the flat-fee way to keep using it: a predictable monthly fee, no per-token math required.
- Standard ($19/user/month with an annual commitment, $22.80 without): higher request quotas, an enterprise SLA, and basic admin controls. The natural step up for Google Cloud-native shops standardizing on Gemini.
- Enterprise ($45/user/month with an annual commitment, $54 without): the highest quotas plus code customization (grounding on your own repos), security review features, and deeper IDE integration.
The advantage of a subscription is predictability: you know your bill in advance and it doesn't spike on a heavy week. The trade-off is that if your usage is spiky, light most weeks, heavy occasionally, you may pay for headroom you rarely use. The plan calculator can help you compare a flat subscription against projected API spend for your actual workload.
Gemini API: Pay-Per-Token Pricing
The Gemini API has no monthly fee: you pay only for the tokens you send (input) and receive (output). This is what you fall back on for production apps, automation, or any usage that outgrows the CLI quotas. List prices on Google's Gemini API pricing page, checked 16 October 2026 (Argon's rates double on December 1, 2026 and Flash's on January 1, 2027; Pro rates are for prompts up to 200K tokens; Argon is not on the free tier):
| Model | Input / M tokens | Output / M tokens | Cached input / M |
|---|---|---|---|
| Gemini 4 Argon | $2.00 ($4.00 from Dec 1) | $10.00 ($20.00 from Dec 1) | $0.10 ($0.20 from Dec 1) |
| Gemini 3.8 Flash | $0.75 ($1.50 from Jan 1) | $3.75 ($7.50 from Jan 1) | $0.075 ($0.15 from Jan 1) |
| Gemini 3.1 Pro Preview | $2.00 | $12.00 | $0.20 |
Two things matter here. First, output is far more expensive than input, roughly 5x on Argon and Flash and 6x on Pro, so chatty, code-heavy generations cost more than long-context reads. Second, cached input is dramatically cheaper (a tenth of the input price on Flash and Pro, a twentieth on Argon), which is why reusing the same large context across turns saves real money. The 1M-token context window is powerful, but a single 500K-token request is not free, it counts against your token bill at the input rate.
SessionWatcher
Know your real Gemini spend.
Stop guessing per-token costs.
Native app for the Mac menu bar and the Windows taskbar. Track Claude, Codex and 20 more tools' usage, costs, and rate limits in real time.
“Fast, simple, and does exactly what it should. Definitely worth it.”
@nicojerome on GitHub
macOS 14+. Free for 7 days, no credit card.
Worked Examples
Token math is abstract until you put real workloads against it. Here are three profiles using the API rates above. (Subscription users on the CLI pay nothing extra for these as long as they stay under quota, these numbers show what the same work would cost on pay-per-token billing.)
Light: casual CLI use on Flash
A developer asking quick questions and doing small edits might burn ~30K tokens/hour, mostly input. Over a 2-hour day that's ~60K tokens. On Gemini 3.8 Flash, assuming a 20% output share, that's about 8 cents per day, roughly $2/month. On a Code Assist license, it is covered by the flat fee.
Moderate: daily IDE + CLI on Pro
A working developer running Gemini Pro 4 hours/day at ~75K tokens/hour lands near 300K tokens/day, or roughly 6.5M tokens/month. At Pro rates with a 20% output share and moderate caching, that's in the range of $20–$26/month on the API. A flat Code Assist Standard subscription ($19/mo with an annual commitment, $22.80 without) is competitive here and removes the quota anxiety.
Heavy: agentic CLI sessions on Pro
Take an agentic, all-day CLI workload of 20M to 40M tokens a month: an illustration, not a measured average, so plug in your own count. On Gemini 3.1 Pro Preview, at the same 20% output share, that can run $80–$160+/month on the API, and more once prompts pass 200K tokens (Pro then bills $4/M input and $18/M output), the point where Enterprise Code Assist or aggressive context caching starts to pay off. This is also where silent 429 lockouts at a daily quota cost you time, not just money.
The Hidden Cost: Lockouts and Bill Surprises
Gemini's pricing has two failure modes that don't show up on the sticker price. On a request quota, the cost is lost time: you hit a 429 mid-task, the CLI retries silently, and you wait for the daily quota to reset (midnight Pacific time on the Gemini API). On the API, the cost is a surprise invoice: token billing is invisible until it arrives, and a few heavy agentic days can multiply your bill.
Both problems have the same root cause, no real-time visibility. Neither AI Studio nor the CLI warns you in the moment. If you can see you're at 80% of your daily quota by lunch, you can pace yourself or switch to Flash for the cheap turns. If you can watch estimated API spend tick up, you never get a shock invoice.
How to Track Your Real Gemini Cost
There are two practical ways to understand what Gemini is actually costing you:
1. The Google Cloud / AI Studio dashboard: shows historical token usage and billing, but with a delay and no live, in-session view. Useful for the monthly reconciliation, useless for pacing yourself today.
2. SessionWatcher: SessionWatcher for Gemini watches your CLI activity live from the Mac menu bar or the Windows taskbar, each model's share of the daily quota with its reset countdown, your pace, tokens and estimated spend today and this month, and notifications at 70% and 90% by default so you can pace before a 429 or a bill spike. It turns Gemini's invisible metering into something you can actually see and control.
Important: that's a different price from Gemini itself.
Google charges you for Gemini (free tier, Code Assist subscription, or API tokens as above). SessionWatcher is a separate, optional app for Mac and Windows that monitors that usage. Gemini monitoring is included with SessionWatcher Pro, $79 one-time or $24/year, which covers all supported tools (Claude Code, Codex, Cursor, Copilot, Gemini, opencode, and more). Every purchase has a 14-day refund.
If your company is buying SessionWatcher for its developers, that is Pro for Teams: $40 per seat per year (or $4 per seat per month), 3-seat minimum, Macs only for now. Each teammate sees their own usage on their own machine; it is not a central team dashboard.
Which Option Is Right for You?
- Solo dev, light-to-moderate use: the free CLI tier is gone, so a paid Gemini API key on Flash is usually the cheapest way to keep using the CLI (the light example above is about $2 a month).
- Steady daily use, predictable budget: Code Assist Standard ($19/mo with an annual commitment, $22.80 without) buys headroom and an SLA without per-token math.
- Production apps or heavy automation: the Gemini API, with caching and Flash-for-cheap-turns to control cost.
- Teams with security/compliance needs: Code Assist Enterprise ($45/mo with an annual commitment, $54 without) for code customization and review features.
Whichever you pick, the only way to make the right call, and to avoid both lockouts and bill surprises, is to know your actual usage. That's exactly what SessionWatcher shows you.
SessionWatcherTrack your real Gemini costs.
Included with Pro. 14-day refund.
Native app for the Mac menu bar and the Windows taskbar. Track Claude, Codex and 20 more tools' usage, costs, and rate limits in real time.
“Fast, simple, and does exactly what it should. Definitely worth it.”
@nicojerome on GitHub
macOS 14+. Free for 7 days, no credit card.
Frequently Asked Questions
How much does Google Gemini cost?
The Gemini app is free with a Google account (Flash-Lite). In the US, Google AI Plus is $4.99 a month, Google AI Pro $19.99 and Google AI Ultra $99.99 (5x Pro's limits) or $199.99 (20x), checked on gemini.google/subscriptions on 16 October 2026; Google lists Gemini 4 Argon on Pro and Ultra. Developers pay the Gemini API per token instead: Gemini 4 Argon is $2 per million input tokens and $10 per million output until November 30, 2026, then $4 and $20.
Is the Gemini CLI free?
No longer for free individual accounts: since June 18, 2026, Gemini CLI does not serve free Gemini Code Assist for individuals or Google AI Pro and Ultra. It still works with a Code Assist Standard or Enterprise license or a paid Gemini API key, which bills per token.
How much does the Gemini API cost per token?
At list prices checked on 16 October 2026: Gemini 4 Argon is $2.00/M input and $10.00/M output until November 30, 2026, then $4.00 and $20.00; Gemini 3.8 Flash is $0.75/M and $3.75/M until December 31, 2026, then $1.50 and $7.50; Gemini 3.1 Pro Preview is $2.00/M input and $12/M output for prompts up to 200K tokens. Cached input is far cheaper ($0.10/M Argon, $0.075/M Flash, $0.20/M Pro). Output dominates cost on generation-heavy work.
How much is Gemini Code Assist Standard?
Standard is $19 per user per month with an annual commitment ($22.80 without), and Enterprise is $45 per user per month with an annual commitment ($54 without). Both add higher quotas, an SLA, and admin controls, and both still serve Gemini CLI now that the free individual tier no longer does. Prices shift, so check the official Gemini pricing page.
How can I track what Gemini actually costs me?
SessionWatcher tracks your Gemini CLI usage live from the Mac menu bar or the Windows taskbar, requests against quota, tokens per minute, and estimated spend. Note this is separate from Google's charges: Gemini monitoring is included with SessionWatcher Pro ($79 one-time or $24/year, all supported tools, 14-day refund), and one license key works on Mac and Windows.