> ## AI Agent Context > - Full documentation index: https://aistudio.google.com/llms.txt > - Gemini Docs MCP server (consistent live info): https://gemini-api-docs-mcp.dev > - Coding agent skills & setup guide: https://aistudio.google.com/docs/coding-agents.md # Understand and count tokens Gemini and other generative AI models process input and output at a granularity called a *token*. **For Gemini models, a token is equivalent to about 4 characters. 100 tokens is equal to about 60-80 English words.** ## About tokens {% id="about-tokens" %} Tokens can be single characters like `z` or whole words like `cat`. Long words are broken up into several tokens. The set of all tokens used by the model is called the vocabulary, and the process of splitting text into tokens is called *tokenization*. When billing is enabled, the [cost of a call to the Gemini API](/docs/pricing) is determined in part by the number of input and output tokens, so knowing how to count tokens can be helpful. You can try out counting tokens in our Colab. [ Try a Colab notebook](https://colab.research.google.com/github/google-gemini/cookbook/blob/main/quickstarts/Counting_Tokens.ipynb) [ View notebook on GitHub](https://colab.research.google.com/github/google-gemini/cookbook/blob/main/quickstarts/Counting_Tokens.ipynb) ## Count tokens {% id="count-tokens" %} All input to and output from the Gemini API is tokenized, including text, image files, and other non-text modalities. You can count tokens in the following ways: * **Call [`count_tokens`](/api-reference/tokens/count-tokens?api-version=v1) with the input of the request.** This returns the total number of tokens in *the input only*. You can make this call before sending the input to the model to check the size of your requests. * **Use the `usage_metadata` attribute on the `response` object after calling `generate_content`.** This returns the total number of tokens in *both the input and the output*: `total_token_count`. It also returns the token counts of the input and output separately: `prompt_token_count` (input tokens) and `candidates_token_count` (output tokens). If you are using a [thinking model](/docs/thinking), the tokens used during the thinking process are returned in `thoughts_token_count`. And if you are using [Context caching](/docs/caching), the cached token count will be in `cached_content_token_count`. ### Count text tokens {% id="text-tokens" %} If you call `count_tokens` with a text-only input, it returns the token count of the text in *the input only* (`total_tokens`). You can make this call before calling `generate_content` to check the size of your requests. Another option is calling `generate_content` and then using the `usage_metadata` attribute on the `response` object to get the following: * The separate token counts of the input (`prompt_token_count`), the cached content (`cached_content_token_count`) and the output (`candidates_token_count`) * The token count for the thinking process (`thoughts_token_count`) * The total number of tokens in *both the input and the output* (`total_token_count`) {% tabs %} {% tab label="Python" %} ```python from google import genai client = genai.Client() prompt = "The quick brown fox jumps over the lazy dog." total_tokens = client.models.count_tokens( model="gemini-4-hydrogen", contents=prompt ) print("total_tokens: ", total_tokens) response = client.models.generate_content( model="gemini-4-hydrogen", contents=prompt ) print(response.usage_metadata) ``` {% /tab %} {% tab label="JavaScript" %} ```js import { GoogleGenAI } from '@google/genai'; const ai = new GoogleGenAI({}); const prompt = "The quick brown fox jumps over the lazy dog."; async function main() { const countTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(countTokensResponse.totalTokens); const generateResponse = await ai.models.generateContent({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(generateResponse.usageMetadata); } await main(); ``` {% /tab %} {% tab label="Go" %} ```go ctx := context.Background() client, err := genai.NewClient(ctx, nil) // Convert prompt to a slice of *genai.Content using the helper. contents := []*genai.Content{ genai.NewContentFromText(prompt, genai.RoleUser), } countResp, err := client.Models.CountTokens(ctx, "gemini-4-hydrogen", contents, nil) if err != nil { return err } fmt.Println("total_tokens:", countResp.TotalTokens) response, err := client.Models.GenerateContent(ctx, "gemini-4-hydrogen", contents, nil) if err != nil { log.Fatal(err) } usageMetadata, err := json.MarshalIndent(response.UsageMetadata, "", " ") if err != nil { log.Fatal(err) } fmt.Println(string(usageMetadata)) ``` {% /tab %} {% /tabs %} ### Count multi-turn (chat) tokens {% id="multi-turn-tokens" %} If you call `count_tokens` with the chat history, it returns the total token count of the text from each role in the chat (`total_tokens`). Another option is calling `send_message` and then using the `usage_metadata` attribute on the `response` object to get the following: * The separate token counts of the input (`prompt_token_count`), the cached content (`cached_content_token_count`) and the output (`candidates_token_count`) * The token count for the thinking process (`thoughts_token_count`) * The total number of tokens in *both the input and the output* (`total_token_count`) To understand how big your next conversational turn will be, you need to append it to the history when you call `count_tokens`. {% tabs %} {% tab label="Python" %} ```python from google import genai from google.genai import types client = genai.Client() chat = client.chats.create( model="gemini-4-hydrogen", history=[ types.Content( role="user", parts=[types.Part(text="Hi my name is Bob")] ), types.Content(role="model", parts=[types.Part(text="Hi Bob!")]), ], ) print( client.models.count_tokens( model="gemini-4-hydrogen", contents=chat.get_history() ) ) response = chat.send_message( message="In one sentence, explain how a computer works to a young child." ) print(response.usage_metadata) extra = types.UserContent( parts=[ types.Part( text="What is the meaning of life?", ) ] ) history = [*chat.get_history(), extra] print(client.models.count_tokens(model="gemini-4-hydrogen", contents=history)) ``` {% /tab %} {% tab label="JavaScript" %} ```js import { GoogleGenAI } from '@google/genai'; const ai = new GoogleGenAI({}); async function main() { const history = [ { role: "user", parts: [{ text: "Hi my name is Bob" }] }, { role: "model", parts: [{ text: "Hi Bob!" }] }, ]; const chat = ai.chats.create({ model: "gemini-4-hydrogen", history: history, }); const countTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: chat.getHistory(), }); console.log(countTokensResponse.totalTokens); const chatResponse = await chat.sendMessage({ message: "In one sentence, explain how a computer works to a young child.", }); console.log(chatResponse.usageMetadata); const extraMessage = { role: "user", parts: [{ text: "What is the meaning of life?" }], }; const combinedHistory = [...chat.getHistory(), extraMessage]; const combinedCountTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: combinedHistory, }); console.log( "Combined history token count:", combinedCountTokensResponse.totalTokens, ); } await main(); ``` {% /tab %} {% tab label="Go" %} ```go ctx := context.Background() client, err := genai.NewClient(ctx, nil) history := []*genai.Content{ {Role: genai.RoleUser, Parts: []*genai.Part({Text: "Hi my name is Bob"})}, {Role: genai.RoleModel, Parts: []*genai.Part({Text: "Hi Bob!"})}, } chat, err := client.Chats.Create(ctx, "gemini-4-hydrogen", nil, history) if err != nil { log.Fatal(err) } firstTokenResp, err := client.Models.CountTokens(ctx, "gemini-4-hydrogen", chat.History(false), nil) if err != nil { log.Fatal(err) } fmt.Println(firstTokenResp.TotalTokens) resp, err := chat.SendMessage(ctx, genai.NewPartFromText("In one sentence, explain how a computer works to a young child.")) if err != nil { log.Fatal(err) } fmt.Printf("%#v\n", resp.UsageMetadata) extra := genai.NewContentFromText("What is the meaning of life?", genai.RoleUser) hist := chat.History(false) hist = append(hist, extra) secondTokenResp, err := client.Models.CountTokens(ctx, "gemini-4-hydrogen", hist, nil) if err != nil { log.Fatal(err) } fmt.Println(secondTokenResp.TotalTokens) ``` {% /tab %} {% /tabs %} ### Count multimodal tokens {% id="multimodal-tokens" %} All input to the Gemini API is tokenized, including text, image files, and other non-text modalities. Note the following high-level key points about tokenization of multimodal input during processing by the Gemini API: * Image inputs with both dimensions <=384 pixels are counted as 258 tokens. Images larger in one or both dimensions are cropped and scaled as needed into tiles of 768x768 pixels, each counted as 258 tokens. * Video and audio files are converted to tokens at the following fixed rates: video at 263 tokens per second and audio at 32 tokens per second. This rate applies to static processing (default). For agentic processing, token usage varies. See [Video token usage by processing mode](/docs/tokens?generate-content=true#video-audio). #### Media resolutions {% id="media-resolutions" %} [Gemini 3 models](/docs/models#gemini-3) introduce granular control over multimodal vision processing with the `media_resolution` parameter. The `media_resolution` parameter determines the **maximum number of tokens allocated per input image or video frame.** Higher resolutions improve the model's ability to read fine text or identify small details, but increase token usage and latency. For more details about the parameter and how it can impact token calculations, see the [media resolution](/docs/media-resolution) guide. #### Image files {% id="images" %} If you call `count_tokens` with a text-and-image input, it returns the combined token count of the text and the image in *the input only* (`total_tokens`). You can make this call before calling `generate_content` to check the size of your requests. You can also optionally call `count_tokens` on the text and the file separately. Another option is calling `generate_content` and then using the `usage_metadata` attribute on the `response` object to get the following: * The separate token counts of the input (`prompt_token_count`), the cached content (`cached_content_token_count`) and the output (`candidates_token_count`) * The token count for the thinking process (`thoughts_token_count`) * The total number of tokens in *both the input and the output* (`total_token_count`) Example that uses an uploaded image from the File API: {% tabs %} {% tab label="Python" %} ```python from google import genai client = genai.Client() prompt = "Tell me about this image" your_image_file = client.files.upload(file=media / "organ.jpg") print( client.models.count_tokens( model="gemini-4-hydrogen", contents=[prompt, your_image_file] ) ) response = client.models.generate_content( model="gemini-4-hydrogen", contents=[prompt, your_image_file] ) print(response.usage_metadata) ``` {% /tab %} {% tab label="JavaScript" %} ```js import { GoogleGenAI } from '@google/genai'; const ai = new GoogleGenAI({}); const prompt = "Tell me about this image"; async function main() { const organ = await ai.files.upload({ file: path.join(media, "organ.jpg"), config: { mimeType: "image/jpeg" }, }); const countTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: createUserContent([ prompt, createPartFromUri(organ.uri, organ.mimeType), ]), }); console.log(countTokensResponse.totalTokens); const generateResponse = await ai.models.generateContent({ model: "gemini-4-hydrogen", contents: createUserContent([ prompt, createPartFromUri(organ.uri, organ.mimeType), ]), }); console.log(generateResponse.usageMetadata); } await main(); ``` {% /tab %} {% tab label="Go" %} ```go ctx := context.Background() client, err := genai.NewClient(ctx, nil) file, err := client.Files.UploadFromPath( ctx, filepath.Join(getMedia(), "organ.jpg"), &genai.UploadFileConfig{ MIMEType : "image/jpeg", }, ) if err != nil { log.Fatal(err) } parts := []*genai.Part{ genai.NewPartFromText("Tell me about this image"), genai.NewPartFromURI(file.URI, file.MIMEType), } contents := []*genai.Content{ genai.NewContentFromParts(parts, genai.RoleUser), } tokenResp, err := client.Models.CountTokens(ctx, "gemini-4-hydrogen", contents, nil) if err != nil { log.Fatal(err) } fmt.Println("Multimodal image token count:", tokenResp.TotalTokens) response, err := client.Models.GenerateContent(ctx, "gemini-4-hydrogen", contents, nil) if err != nil { log.Fatal(err) } usageMetadata, err := json.MarshalIndent(response.UsageMetadata, "", " ") if err != nil { log.Fatal(err) } fmt.Println(string(usageMetadata)) ``` {% /tab %} {% /tabs %} Example that provides the image as inline data: {% tabs %} {% tab label="Python" %} ```python from google import genai import PIL.Image client = genai.Client() prompt = "Tell me about this image" your_image_file = PIL.Image.open(media / "organ.jpg") print( client.models.count_tokens( model="gemini-4-hydrogen", contents=[prompt, your_image_file] ) ) response = client.models.generate_content( model="gemini-4-hydrogen", contents=[prompt, your_image_file] ) print(response.usage_metadata) ``` {% /tab %} {% tab label="JavaScript" %} ```js import { GoogleGenAI } from '@google/genai'; const ai = new GoogleGenAI({}); const prompt = "The quick brown fox jumps over the lazy dog."; async function main() { const countTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(countTokensResponse.totalTokens); const generateResponse = await ai.models.generateContent({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(generateResponse.usageMetadata); } await main(); ``` {% /tab %} {% tab label="Go" %} ```js import { GoogleGenAI } from '@google/genai'; const ai = new GoogleGenAI({}); const prompt = "The quick brown fox jumps over the lazy dog."; async function main() { const countTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(countTokensResponse.totalTokens); const generateResponse = await ai.models.generateContent({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(generateResponse.usageMetadata); } await main(); ``` {% /tab %} {% /tabs %} #### Video or audio files {% id="video-audio" %} Token calculation for video depends on the processing mode: | **Processing mode** | **Frames sampled** | **Tokens per second** | **Tokens for 60s video** | | --- | --- | --- | --- | | **Static** (default) | 1 FPS (all frames) | ~100 (default low resolution) / ~300 (high resolution) | ~6,000 (default) / ~18,000 (high) | | **Agentic** | Dynamic (on demand) | Variable (up to 88% fewer) | Typically 1,500 to 3,000 | For details on setting the processing mode, see [Agentic video understanding](/docs/video-understanding#agentic-video-understanding). Audio and video are each converted to tokens at the following fixed rates: * Video: 263 tokens per second (static mode) * Audio: 32 tokens per second If you call `count_tokens` with a text-and-video/audio input, it returns the combined token count of the text and the video/audio file in *the input only* (`total_tokens`). You can make this call before calling `generate_content` to check the size of your requests. You can also optionally call `count_tokens` on the text and the file separately. Another option is calling `generate_content` and then using the `usage_metadata` attribute on the `response` object to check token usage. In `UsageMetadata`, agentic video tokens are reported across the following fields: * **Initial prompt** (video reference + user prompt): `prompt_token_count` * **Navigation thinking**: `thoughts_token_count` * **Transcript, frames, and audio loaded on demand**: `tool_use_prompt_token_count` * **Final answer**: `candidates_token_count` {% tabs %} {% tab label="Python" %} ```js import { GoogleGenAI } from '@google/genai'; const ai = new GoogleGenAI({}); const prompt = "The quick brown fox jumps over the lazy dog."; async function main() { const countTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(countTokensResponse.totalTokens); const generateResponse = await ai.models.generateContent({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(generateResponse.usageMetadata); } await main(); ``` {% /tab %} {% tab label="JavaScript" %} ```js import { GoogleGenAI } from '@google/genai'; const ai = new GoogleGenAI({}); const prompt = "The quick brown fox jumps over the lazy dog."; async function main() { const countTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(countTokensResponse.totalTokens); const generateResponse = await ai.models.generateContent({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(generateResponse.usageMetadata); } await main(); ``` {% /tab %} {% tab label="Go" %} ```js import { GoogleGenAI } from '@google/genai'; const ai = new GoogleGenAI({}); const prompt = "The quick brown fox jumps over the lazy dog."; async function main() { const countTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(countTokensResponse.totalTokens); const generateResponse = await ai.models.generateContent({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(generateResponse.usageMetadata); } await main(); ``` {% /tab %} {% /tabs %} ### Count thought tokens {% id="thought-tokens" %} {% callout type="note" %} The API provides **summaries** in both the [free and paid tiers](/docs/pricing). **Thought signatures** increase the input token count (and your cost) when you send them back in subsequent conversational turns. {% /callout %} When you turn on thinking, response pricing is the sum of output tokens and thinking tokens. You can retrieve the total number of generated thinking tokens from the `thoughtsTokenCount` field (or SDK equivalent). {% tabs %} {% tab label="Python" %} ```js import { GoogleGenAI } from '@google/genai'; const ai = new GoogleGenAI({}); const prompt = "The quick brown fox jumps over the lazy dog."; async function main() { const countTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(countTokensResponse.totalTokens); const generateResponse = await ai.models.generateContent({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(generateResponse.usageMetadata); } await main(); ``` {% /tab %} {% tab label="JavaScript" %} ```js import { GoogleGenAI } from '@google/genai'; const ai = new GoogleGenAI({}); const prompt = "The quick brown fox jumps over the lazy dog."; async function main() { const countTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(countTokensResponse.totalTokens); const generateResponse = await ai.models.generateContent({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(generateResponse.usageMetadata); } await main(); ``` {% /tab %} {% tab label="Go" %} ```js import { GoogleGenAI } from '@google/genai'; const ai = new GoogleGenAI({}); const prompt = "The quick brown fox jumps over the lazy dog."; async function main() { const countTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(countTokensResponse.totalTokens); const generateResponse = await ai.models.generateContent({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(generateResponse.usageMetadata); } await main(); ``` {% /tab %} {% /tabs %} Thinking models generate full thoughts to improve the quality of the final response, and then output [summaries](/docs/thinking#summaries) to provide insight into the thought process. So, the API bases pricing on the full thought tokens the model generates to create a summary, even though the API only outputs the summary. You can learn more about how to configure thinking in the [Gemini thinking](/docs/thinking) guide. ## Context windows {% id="context-windows" %} The models available through the Gemini API have context windows that are measured in tokens. The context window defines how much input you can provide and how much output the model can generate. You can determine the size of the context window by calling the [`models.get` endpoint](/api-reference/models/get-model?api-version=v1) or by looking in the [models documentation](/docs/models). * **Gemini 4 models (`gemini-4-carbon`, `gemini-4-argon`, and `gemini-4-hydrogen`):** Share a single 1M-token context window (`1,048,576` tokens) across input, thinking, and output tokens. Single requests are capped at `65,536` output tokens on `gemini-4-hydrogen` (streaming and non-streaming) and on streaming (`streamGenerateContent`) for `gemini-4-carbon` and `gemini-4-argon` (`32,768` output tokens on direct non-streaming `generateContent` for Carbon and Argon), or you can generate up to `1,048,576` tokens using [long decode](/docs/long-decode?generate-content=true). If your input tokens plus `max_output_tokens` exceed 1,048,576 tokens, the API automatically caps the output limit to the remaining space in the context window (`1,048,576` minus your input tokens). * **Gemini 3 and 2.5 models:** Use separate input (`1,048,576` tokens) and output (`65,536` tokens) budgets. {% tabs %} {% tab label="Python" %} ```js import { GoogleGenAI } from '@google/genai'; const ai = new GoogleGenAI({}); const prompt = "The quick brown fox jumps over the lazy dog."; async function main() { const countTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(countTokensResponse.totalTokens); const generateResponse = await ai.models.generateContent({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(generateResponse.usageMetadata); } await main(); ``` {% /tab %} {% tab label="JavaScript" %} ```js import { GoogleGenAI } from '@google/genai'; const ai = new GoogleGenAI({}); const prompt = "The quick brown fox jumps over the lazy dog."; async function main() { const countTokensResponse = await ai.models.countTokens({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(countTokensResponse.totalTokens); const generateResponse = await ai.models.generateContent({ model: "gemini-4-hydrogen", contents: prompt, }); console.log(generateResponse.usageMetadata); } await main(); ``` {% /tab %} {% tab label="Go" %} ```go ctx := context.Background() client, err := genai.NewClient(ctx, nil) // Convert prompt to a slice of *genai.Content using the helper. contents := []*genai.Content{ genai.NewContentFromText(prompt, genai.RoleUser), } countResp, err := client.Models.CountTokens(ctx, "gemini-4-hydrogen", contents, nil) if err != nil { return err } fmt.Println("total_tokens:", countResp.TotalTokens) response, err := client.Models.GenerateContent(ctx, "gemini-4-hydrogen", contents, nil) if err != nil { log.Fatal(err) } usageMetadata, err := json.MarshalIndent(response.UsageMetadata, "", " ") if err != nil { log.Fatal(err) } fmt.Println(string(usageMetadata)) ``` {% /tab %} {% /tabs %}