Daily request quota
The free tier includes a requests-per-day quota that resets at midnight Pacific time. It is enough for prototyping and low-volume side projects, but AI agents and coding tools can exhaust it quickly.
Claude visitors: new accounts receive $1.00 in bonus API credit within 30 days. One per person; usage credits only.
LLM API Pricing
Gemini API Pricing in 2026
Google bills Gemini per million tokens and offers a rate-limited free tier for prototyping. The current lineup is Gemini 3.x Flash and Gemini 3.1 Pro Preview; Google now limits Gemini 2.5 to users who have used it before. RunAPI serves these models through one OpenAI-compatible key, and the table below lists official and RunAPI rates.
Updated October 01, 2026
RunAPI Editorial
At a glance
What does the Gemini API cost right now?
Google publishes per-token prices for each Gemini model. The current lineup has Gemini 3.x Flash models for fast, high-volume tasks and Gemini 3.1 Pro Preview for complex reasoning. Google limits Gemini 2.5 Pro and 2.5 Flash to users who have used them before and points new projects to newer models, so this page marks them as legacy. All prices below are per million tokens, the billing unit Google uses.
Fast, high-volume tasks
Gemini 3.x Flash — the current Flash models for high-volume and latency-sensitive work.
Maximum capability
Gemini 3.1 Pro Preview — the current Pro model for complex reasoning.
Free tier
Google's free tier has per-minute and per-day limits that vary by model; your current values are shown in Google AI Studio. It is enough for prototyping, but an AI agent can exhaust it quickly.
Compare flagships
The comparison table puts Gemini 3.1 Pro Preview next to Claude Sonnet and GPT-5.4 at current prices.
Model-by-model breakdown
How much does each Gemini model cost per million tokens?
The table below shows official Google pricing alongside RunAPI pricing from the current price list. Rows marked Legacy are Gemini 2.5 models, which Google now limits to users who have used them before. No volume commits, no subscriptions.
| Model | Official input /M | Official output /M | RunAPI input /M | RunAPI output /M |
|---|---|---|---|---|
| gemini-3.8-flash | $0.75 | $3.75 | $0.60 20% OFF | $3.00 |
| gemini-3.7-flash | $0.75 | $3.75 | $0.60 20% OFF | $3.00 |
| gemini-3.6-flash | $0.75 | $3.75 | $0.60 20% OFF | $3.00 |
| gemini-3.5-flash | $1.50 | $9.00 | $1.20 20% OFF | $7.20 |
| gemini-3.5-flash-lite | $0.30 | $2.50 | $0.24 20% OFF | $2.00 |
| gemini-3.1-pro-preview Higher rate on long prompts | $2.00 | $12.00 | $1.60 20% OFF | $9.60 |
| gemini-2.5-pro Higher rate on long prompts Legacy | $1.25 | $10.00 | $1.00 20% OFF | $8.00 |
| gemini-2.5-flash Legacy | $0.30 | $2.50 | $0.24 20% OFF | $2.00 |
Free tier
How does Google's Gemini free tier work?
Google offers a free tier so you can prototype before paying. Free-tier usage is capped by requests per minute, tokens per minute, and requests per day; the values depend on the model, and Google lists your current limits in Google AI Studio. When you exceed a limit, requests return a rate limit error. The Gemini API and Google AI Studio are only available in certain countries and territories; if they are not available in your region, RunAPI is the next step.
Rate limits
Per-minute limits on requests and tokens apply per project, not per API key, and Google notes that they can change. Burst traffic — like running an AI agent — can hit these limits fast and start returning errors.
Data usage terms
Free-tier inputs may be used to improve Google's products. Paid-tier and RunAPI traffic is not used for training, which matters for sensitive data.
When to upgrade
Move to paid or RunAPI once you need steady throughput, higher rate limits, or stronger data handling. RunAPI bills per token with no daily cap.
Provider comparison
Is Gemini cheaper than Claude and GPT?
Developers often compare Gemini against Claude Sonnet and GPT-5.4. Here is how the flagship models stack up per million tokens at current prices.
| Provider | Flagship model | Input /M | Output /M | RunAPI rate |
|---|---|---|---|---|
| gemini-3.1-pro-preview | $2.00 | $12.00 | $1.60 / $9.60 | |
| OpenAI | gpt-5.4 | $2.50 | $15.00 | $2.00 / $12.00 |
| Anthropic | claude-sonnet-4-6 | $3.00 | $15.00 | $2.40 / $12.00 |
Getting started
How to access the Gemini API through RunAPI
1
Create a RunAPI account
Sign up and generate an API key from the dashboard.
2
Copy your API key
Go to Dashboard → API Keys. Create a key and save it — you will use this as your OpenAI API key.
3
Point your SDK to RunAPI
Set the base URL to https://runapi.ai/v1 and use your RunAPI API key. Any OpenAI-compatible client works — Python, Node.js, Go, Ruby, or curl.
4
Start making requests
Use gemini-3.8-flash, gemini-3.5-flash-lite, gemini-3.1-pro-preview, or another Gemini model ID from the table in the model parameter. RunAPI handles routing and bills at the rates listed on this page.
Frequently asked questions
Gemini API Pricing FAQ
How much does the Gemini API cost?
The pricing table above lists Google's official rate and the RunAPI rate per million tokens for Gemini 3.x Flash models and Gemini 3.1 Pro Preview, plus the legacy Gemini 2.5 models, from the current price list. There is no subscription or volume commitment; you pay only for the tokens each request uses.
Is Gemini cheaper than Claude and GPT?
Compare the input and output rates in the provider comparison table above. They come from the current price list for all three providers, so the comparison reflects today's prices.
What are the rate limits on the free tier, and what happens when I hit them?
Free-tier usage is capped by requests per minute, tokens per minute, and requests per day. The values depend on the model and your usage tier; Google shows your current limits in Google AI Studio and documents how they work at ai.google.dev/gemini-api/docs/rate-limits. When you exceed a limit, requests return a rate limit error. Free-tier inputs may be used to improve Google's products, so avoid sensitive data. Once you need steady throughput, move to paid.
Can I use Gemini with the OpenAI SDK?
Yes. RunAPI is OpenAI-compatible. Point any OpenAI client at runapi.ai/v1 and use Gemini model IDs like gemini-3.6-flash or gemini-3.1-pro-preview. Existing OpenAI SDK code works without changes beyond the base URL and key.
Can I still use Gemini 2.5 Pro and 2.5 Flash?
Google now limits Gemini 2.5 access to users who have used those models before and recommends newer models for new projects. The table keeps the 2.5 rows, marked Legacy, for existing workloads. For a new project, Google recommends Gemini 3.5 Flash-Lite or Gemini 3.8 Flash; both are available on RunAPI.
I keep running out of free tier limits. What are my options?
RunAPI gives you one API key that works across Gemini, Claude, and GPT — no Google Cloud project setup, no billing account to manage. If you have hit free-tier limits or the Gemini API is not available in your region, RunAPI's paid access is available globally.
How does billing work?
Pay-as-you-go. You fund your account with a balance, and each API call deducts the token cost at the rates listed on this page. No subscriptions, no invoices, no contracts. You can monitor spend per key from the RunAPI dashboard and set up alerts before the balance runs low.
RELATED
Explore More
Claude API Pricing
Claude Opus and Sonnet pricing with official and RunAPI rates.
OpenAI API Pricing
The full GPT lineup with official and RunAPI rates.
Windsurf Pricing
What Windsurf costs with its own credits vs API billing.
Model Catalog
Browse all 250+ models with pricing, parameters, and code samples.
Start using Gemini with one API key.
Create a free RunAPI account, get your API key, and start calling Gemini 3.x Flash or Gemini 3.1 Pro Preview at the rates listed on this page.