LATESTSAAS FOR PEOPLE WHO BUILDAI · SOFTWARE · CREATORS
AI BENCHMARK · PROVIDER

Google Gemini: models, prices and how to use them

Gemini is Google’s family of AI models. It is used through the Gemini app, through Google products and through the API. In AI Benchmark we track 9 models on Google with prices, context, speed and usage rates. The highest Intelligence Index belongs to Gemini 4 Neon (61,4). The data is updated every day.

Prices are in dollars per 1M tokens: from $0 for вход at Gemma 4 31B up to the more powerful models. The largest context window is 1 million tokens (Gemini 4 Neon), that is, how much text the model holds in a single conversation. Fastest is Gemini 3.5 Flash-Lite with about 407 tokens per second. Below are all the reasoning levels (effort) of each model as separate rows: Intelligence Index score, cost per task, input and output price, cache, time to first token and speed. The Intelligence Index is an overall score from independent Artificial Analysis tests. A higher reasoning level usually gives a better result, but uses more tokens and costs more.

Tracked models9with a separate page
Highest Intelligence Index61,4Gemini 4 Neon
Lowest input price$0per 1M tokens, Gemma 4 31B
Largest context1 milliontokens, Gemini 4 Neon
Fastest model407tokens/sec., Gemini 3.5 Flash-Lite

Which model from Google to choose

The choice depends directly on the numbers above. For your task, test with real examples.

All tracked models from Google

Gemini 4 Neon

Intelligence Index61,4
Input / 1M$3
Output / 1M$15
Context1 million
Speed136 t/s
10,000 requests*$105

Gemini 4 Neon is Google’s model for software development, knowledge work and science, announced on 18 November 2026. Its strongest level (Xhigh) has an Intelligence Index of 61.4 and ranks #3 of 84 tracked models. The API price is $3 per 1M input and $15 for 1 million…

Pricing and how it is usedReleased: 18 November 2026

Gemini 4 Carbon

Intelligence Index60
Input / 1M$4
Output / 1M$20
Contextn/a
Speed82 t/s
10,000 requests*$140

Gemini 4 Carbon is a Google model published in Artificial Analysis data dated October 2026. Its top level (High) has an Intelligence Index of 60 and ranks 3rd out of 82 tracked models. The API price is $4 per 1 million input and $20 for 1 million andзходящand т…

Pricing and how it is usedReleased: October 2026

Gemini 3.8 Flash

Intelligence Index30,8
Input / 1M$0,75
Output / 1M$3,75
Contextn/a
Speed124 t/s
10,000 requests*$26,25

Gemini 3.8 Flash is a Google model listed in Artificial Analysis data dated September 2026. Its strongest level (High) has an Intelligence Index of 30.8 and ranks 29th out of 82 tracked models. The API price is $0.75 per 1M input and $3,75 for 1 million …

Pricing and how it is usedReleased: September 2026

Gemini 3.5 Flash-Lite

Intelligence Index11,7
Input / 1M$0,3
Output / 1M$2,5
Contextn/a
Speed407 t/s
10,000 requests*$15,5

Gemini 3.5 Flash-Lite is a Google model published in Artificial Analysis data dated July 2026. Its top tier has an Intelligence Index of 11.7 and ranks 61st of 82 tracked models. The API price is $0.3 per 1M input and $2,5 for 1 million andзходящand т…

Pricing and how it is usedReleased: July 2026

Gemini 3.1 Flash-Lite

Intelligence Index6,2
Input / 1M$0,25
Output / 1M$1,5
Contextn/a
Speed325 t/s
10,000 requests*$10

Gemini 3.1 Flash-Lite is a Google model, listed in Artificial Analysis data dated March 2026. Its top level has an Intelligence Index of 6.2 and ranks 69th out of 82 tracked models. The API price is $0.25 per 1M input and $1,5 for 1 million andзходящand …

Pricing and how it is usedReleased: March 2026

Gemma 4 E4B

Intelligence Index4,8
Input / 1M$0,02
Output / 1M$0,1
Contextn/a
Speed70 t/s
10,000 requests*$0,7

Gemma 4 E4B is a Google model, published in Artificial Analysis data dated April 2026. Its strongest tier (Reasoning) has an Intelligence Index of 4.8 and ranks 73rd out of 82 tracked models. The API price is $0.02 per 1M input and $0,1 for 1 million andзходя…

Pricing and how it is usedReleased: April 2026

Gemma 4 31B

Intelligence Index4,4
Input / 1M$0
Output / 1M$0
Contextn/a
Speed35 t/s
10,000 requests*$0

Gemma 4 31B is a Google model, published in Artificial Analysis data dated April 2026. Its strongest level (Reasoning) has an Intelligence Index of 4.4 and ranks 76th out of 82 tracked models. The API price is $0 per 1M input and $0 for 1 million andзходящand то…

Pricing and how it is usedReleased: April 2026

Gemini 4 Argon

Intelligence Indexn/a
Input / 1M$2
Output / 1M$10
Context1 million
Speedn/a
10,000 requests*$70

Gemini 4 Argon is Google’s frontier model, announced on 13 October 2026. It is available through the Gemini API, AI Studio and Google apps. It has a combined limit of 1M tokens for input and output.

Pricing and how it is usedReleased: October 2026

Gemini 4 Hydrogen

Intelligence Indexn/a
Input / 1M$1,5
Output / 1M$7,5
Context1 million
Speedn/a
10,000 requests*$52,5

Gemini 4 Hydrogen is a Google model for running agentic workflows, announced on November 12, 2026. The price is the same across the whole 1M token context window. Google positions it as the fast model, with Carbon and Argon as the deep reasoning models.

Pricing and how it is usedReleased: November 2026

* Example API spend for 10,000 requests of 1000 input and 500 output tokens each. The actual bill depends on the length of the texts.

How to use Gemini

Gemini 4 Neon has been available since 18 November 2026 in the Gemini API (Google AI Studio and the paid tier), in Gemini for business, in Google Antigravity and in the Gemini app for Google AI Pro and Ultra subscribers.

  1. For chat: open the Gemini app and select ’In-depth analysis: 4 Neon" from the dropdown menu (for Pro and Ultra subscribers).
  2. For the API: sign in to Google AI Studio with a Google account and create a key. The model is paid.
  3. In the request, specify the model gemini-4-neon. It works through generateContent and the Interactions API.
  4. The reasoning levels are Low, Medium, High (default) and Xhigh. For most tasks start with High: Xhigh costs about 1.7 times more per task for an index less than 2 points higher. Check the exact parameter name in the Gemini API documentation.
  5. For text in languages other than English, keep in mind that tokens are more numerous than in English, and the same text ends up costing more. Check the quality of the answers on your own real examples.

The rates are from the page of Gemini 4 Neon. For the other models the same channels apply, and the model name is changed in the request.

Compare models with each other

Compare Gemini models side by side

Frequently asked questions about Gemini

Which models does Google tracks AI Benchmark?

Currently we track 9 models: Gemini 4 Neon, Gemini 4 Carbon, Gemini 3.8 Flash, Gemini 3.5 Flash-Lite, Gemini 3.1 Flash-Lite, Gemma 4 E4B, Gemma 4 31B, Gemini 4 Argon, Gemini 4 Hydrogen. Each has its own page with price, context and usage rates. The list and data are updated every day.

Which is the best model for Google?

Among the tracked models, the highest Artificial Analysis Intelligence Index belongs to Gemini 4 Neon: 61,4 points. The index is an overall score from several independent tests and does not say which model is best for your specific task.

How much it costs Gemini via API?

Prices are per 1M tokens. The cheapest model is Gemma 4 31B: $0 for вход and $0 for andзход. The highest output price is $20. Monthly app subscriptions are not included in these numbers.

Which model from Google is the fastest?

Gemini 3.5 Flash-Lite responds fastest among those tracked: about 407 tokens per second according to Artificial Analysis data.

How large is the context of Gemini?

The largest context window is 1 million tokens at Gemini 4 Neon. This is how much text the model can hold in one conversation.

Are open-weight models the models from Google?

None of the tracked models have open weights. They are used through apps and APIs.

How to use Gemini?

Gemini 4 Neon has been available since 18 November 2026 in the Gemini API (Google AI Studio and the paid tier), in Gemini for business, in Google Antigravity and in the Gemini app for Google AI Pro and Ultra subscribers.

Sources: Artificial Analysis, the official pages of Google. Last check: 22.11.2026. To the full AI Benchmark

FOR CREATORS

Give your product a spot.

Submit the product for editorial review. It will be published after approval.

Submit only public product information. How we handle requests.