Google's Gemini API is still the most generous first-party free tier among the big labs, and Gemini 3.8 Flash is the model most developers reach for first on it. This guide explains what the free tier includes, what it costs to leave it, and the one date to put in your calendar: Google has published that paid prices for Gemini 3.8 Flash double on January 1, 2027. All prices come from Google's Gemini API pricing page, and availability from the Gemini 3.8 Flash page in our catalog, checked on November 12 and 13, 2026.
What Gemini 3.8 Flash is
Google describes Gemini 3.8 Flash as its most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows. In our catalog it is a mid-size, proprietary model with a 1M-token context window that accepts images, supports step-by-step reasoning, coding, and tool calling. It has been free in the API since October 14, 2026, according to our records.
| Artificial Analysis metric | Gemini 3.8 Flash (high) |
|---|---|
| Intelligence Index | 40.9 |
| GPQA Diamond | 95.3% |
| Humanity's Last Exam | 47.8% |
| Long-context reasoning (LCR) | 81.3% |
These are third-party figures. On GPQA Diamond and Humanity's Last Exam the model scores higher than several larger open-weight models in our catalog, while its overall index sits close to them, so the right comparison depends on the task. The best free coding models guide lists the alternatives.
What the free tier includes
According to Google's pricing page, Gemini 3.8 Flash on the Free tier is free of charge for input tokens, output tokens (including thinking tokens), and context caching. Several things are not included or work differently:
- Grounding with Google Search and Maps is not available on the free tier. On the paid tier, Google includes 5,000 free search requests per month shared across all Gemini 3 and newer models, then charges $14 per 1,000 requests, with a daily cap.
- Your content may be used to improve Google's products on the free tier. On the paid tier it is not. This matters if you send private code or customer data.
- Rate limits apply per project and vary by model. Google directs developers to AI Studio for the exact current limits, and daily quotas reset at midnight Pacific time.
Models differ in whether they have a free tier at all. In our catalog's record of the Gemini API free tier, free models include Gemini 4 Hydrogen, 3.8 Flash, 3.6 Flash, 3.5 Flash-Lite, 3.1 Flash-Lite, 3 Flash Preview, 2.5 Pro, 2.5 Flash, 2.5 Flash-Lite, and Gemma 4. Gemini 4 Carbon, Gemini 4 Argon, and 3.1 Pro Preview are paid only. Google's pricing page repeats this for Carbon and Argon: neither is available on the Free tier, and every token they process is billed, including thinking tokens. See the credits page for the full free-tier entry and its last verification date.
What leaving the free tier costs
Prices per 1M tokens for the paid tier, from Google's pricing page:
| Item | Through Dec 31, 2026 | From Jan 1, 2027 |
|---|---|---|
| Gemini 3.8 Flash input | $0.75 | $1.50 |
| Gemini 3.8 Flash output (including thinking) | $3.75 | $7.50 |
| Context caching | $0.075 | $0.15 |
| Cache storage per 1M tokens per hour | $0.50 | $1.00 |
Google lists the higher prices as taking effect January 1, 2027. If you are budgeting for next year, use the January numbers. For comparison, Google's page lists Gemini 4 Hydrogen at $1.50 input and $7.50 output per 1M tokens on the paid tier, with a free tier as well, and Gemini 4 Argon and Carbon at $2.00 input and $10.00 output with no free tier.
A cost comparison
For a job with 1M input tokens and 200K output tokens:
- Gemini 3.8 Flash today: 1M × $0.75 + 0.2M × $3.75 = $0.75 + $0.75 = $1.50
- Gemini 3.8 Flash from January 1, 2027: 1M × $1.50 + 0.2M × $7.50 = $3.00
- The same job on free routes for open models such as DeepSeek V4.1 Flash costs nothing on the free routes, or about $0.18 to $0.37 at DeepSeek's own prices for a typical cached agent task.
So on pure price, Gemini 3.8 Flash is not the cheapest paid option, and it becomes more expensive in January. Its strengths are the free tier, the 1M context, image input, and strong scores on the reasoning benchmarks.
Calling it with the OpenAI SDK
Google exposes an OpenAI-compatible endpoint, so existing code needs only a base URL and key change.
from openai import OpenAI
client = OpenAI(
base_url="https://generativelanguage.googleapis.com/v1beta/openai/",
api_key="YOUR_GEMINI_API_KEY",
)
reply = client.chat.completions.create(
model="gemini-3.8-flash",
messages=[{"role": "user", "content": "Summarize this changelog."}],
)
print(reply.choices[0].message.content)Create the key in Google AI Studio. No card is needed for the free tier, according to our catalog entry. The same call works from Cursor, Cline, OpenCode, and Open WebUI by changing the provider settings.
Free tier strategy
- Treat the free tier as development capacity. It is excellent for prototypes, evaluations, and small internal tools. It is not an SLA.
- Keep sensitive data out. Because free-tier content may be used to improve Google products, do not send customer data or proprietary code through it. Use the paid tier for that.
- Add a fallback model. Handle HTTP 429 with exponential backoff and route to a second provider. The free LLM API guide explains how to stack providers.
- Plan the January price change. If you move to the paid tier, model the new rates now.
- Cache aggressively. Context caching is priced far below fresh input, and on the free tier it costs nothing.
Other Google free options worth knowing
- Google Antigravity has a $0 plan with agent models, unlimited Tab completions and Command requests, and basic weekly limits. Gemini CLI stopped serving free Google-login users on June 18, 2026, and Antigravity CLI replaces it, according to Google's announcement and the Gemini CLI documentation.
- Jules, Google's coding agent, has a free plan with 15 tasks per rolling 24 hours and 3 concurrent tasks.
- Google AI Pro for students is free for one year for eligible U.S. college students who redeem by December 31, 2026, after which it renews at $19.99 per month. See our student perks guide.
FAQ
Is Gemini 3.8 Flash free?
Yes, on the Gemini API Free tier, input and output tokens are free of charge, with rate limits per project. Paid-tier usage is billed.
How much will Gemini 3.8 Flash cost in 2027?
Google lists $1.50 per 1M input tokens and $7.50 per 1M output tokens from January 1, 2027, up from $0.75 and $3.75 through December 31, 2026.
Does the free tier use my data for training?
Google's pricing page says content on the free tier may be used to improve its products, and that paid-tier content is not.
Can I use Google Search grounding for free?
No. Grounding is not available on the free tier. On the paid tier, 5,000 requests per month are free across Gemini 3 and newer models, then $14 per 1,000.
Which Gemini models have no free tier?
Gemini 4 Carbon and Gemini 4 Argon are not available on the free tier, according to Google's pricing page.
What are the exact rate limits?
They vary by model and project, and Google publishes the current values in AI Studio and the Gemini API rate-limits documentation. We do not republish numbers that change often.
