Prices are USD per 1 million tokens as shown on Google's pricing page on 19 November 2026. The page lists many models and scheduled changes, so verify the current figures before you budget.
What are the Gemini API pricing tiers?
- Free tier: limited model access with free input and output tokens. Google may use your content to improve its products.
- Paid tier: higher rate limits, context caching and the Batch API at a 50% cost reduction. Your content is not used to improve Google products.
- Enterprise tier: available by contacting Google sales.
How much does the Gemini API cost per million tokens?
Standard prices for a selection of text, multimodal and reasoning models (input / output, USD per 1M tokens):
| Model | Free tier | Standard input / output |
|---|---|---|
| Gemini 4 Carbon and Argon | No | $2.00 / $10.00, rising to $4.00 / $20.00 from 1 Dec 2026 |
| Gemini 4 Neon | No | $3.00 / $15.00 |
| Gemini 4 Hydrogen | Yes | $1.50 / $7.50 |
| Gemini 3.8 Flash | Yes | $0.75 / $3.75, rising to $1.50 / $7.50 from 1 Jan 2027 |
| Gemini 3.5 Flash-Lite | Yes | $0.30 / $2.50 |
| Gemini 3.1 Pro Preview | No | $2.00 / $12.00 up to 200k tokens; $4.00 / $18.00 above |
| Gemini 2.5 Flash-Lite | Yes | $0.10 / $0.40 (text, image, video input) |
Batch and Flex pricing is about half of Standard. Priority pricing costs more. Context caching has its own rates, and some models support implicit caching only.
What other Gemini API features are priced?
- Search grounding: 5,000 grounded prompts per month are free and shared across Gemini 3 and later models, then USD 14 per 1,000, with a cap of 1,500 per day.
- Live and speech models: priced per million tokens or per minute of audio, with Free tier listed for several of them.
- Image and video generation: Gemini Omni 1.2 Flash video is about USD 0.07 per second at 720p. Lyria 3.5 music is USD 0.08 per full song.
What Gemini API price changes are coming?
- 1 December 2026: Gemini 4 Carbon and Argon Standard prices roughly double. Requests to gemini-2.5-pro route to Gemini 3.8 Flash, and gemini-2.5-flash routes to Gemini 3.5 Flash-Lite.
- 3 December 2026: Gemini Omni 1.1 requests route to Omni 1.2.
- 1 January 2027: Gemini 3.8 Flash, 3.6 Flash and the 3.8 text-to-speech models roughly double in price.
- 15 March 2027: Gemini 2.5 Flash Image shuts down.
How can you keep Gemini API costs down?
- Use the Batch API for work that can wait, for a 50% saving.
- Match the model to the task: Flash-Lite class models for classification and extraction, larger models only where quality needs it.
- Use context caching for long, repeated prompts.
- Track grounding use, since Search grounding is billed after the free monthly allowance.
- Set budgets and alerts before the December and January price changes land.
Building with Gemini and other LLMs
Costs depend on prompts as much as on price lists. Emerald Global trains teams in Bangkok and online on Generative AI, AI agents and LLM tools including Gemini, ChatGPT, Claude and Microsoft Copilot, in English or Thai, so your team spends tokens on work that matters.
Frequently asked questions
Is the Gemini API free?
Google lists a Free tier with limited model access and free input and output tokens. Google may use content from the Free tier to improve its products. The Paid tier does not.
How much does the Gemini API cost?
Paid pricing is per 1 million tokens and varies by model. For example, Google lists Gemini 4 Carbon at USD 2.00 input and USD 10.00 output, rising to USD 4.00 and USD 20.00 from 1 December 2026.
What is the Gemini Batch API discount?
Google lists Batch pricing at about 50% below Standard on the Paid tier.
When do Gemini API prices change?
Google lists increases for Gemini 4 Carbon and Argon on 1 December 2026, and for Gemini 3.8 Flash, 3.6 Flash and the 3.8 text-to-speech models on 1 January 2027.
Does Emerald Global teach Gemini and other AI tools?
Yes. Emerald Global in Bangkok trains teams on Generative AI, AI agents and LLM tools, on-site or online, in English or Thai.
Sources
Figures come from the pages above, read on 19 November 2026. Prices and plans change, so check the official page before you decide.