A token is a piece, not a word.
Tokens can represent words, parts of words, spaces, punctuation, or bytes. Language and the selected tokenizer change the count. Try the multilingual example above.
Token field guide
Make every
token count.
A hands-on counter. A current model directory. A clearer idea of what you are paying for.
Try it
Small text. Surprising numbers.
Local text estimate
—
Using o200k_base. Message framing is extra.
- Characters
- 12
- Words (space separated)
- 2
- Estimated text input cost
- Unavailable
Only this button sends your text to OpenAI. Local counts stay in your browser.
The text model directory
Fresh data. Fewer guesses.
Context is the total working window. Maximum output is a separate limit. Prices below are standard USD per million tokens.
73 models
| Model | Context | Max output | Input / 1M | Output / 1M | Capabilities |
|---|---|---|---|---|---|
GPT-6 Astraopenai · gpt-6-astraLong context pricingAbove 272,000 input tokens: $20 input / $75 output per 1M | 1,050,000 | 128,000 | $10$1 cached | $50 | Vision · Reasoning |
| Gemini 3.8 Flashgoogle · gemini-3.8-flash | 1,048,576 | 65,536 | $0.75$0.075 cached | $3.75 | Vision · Reasoning |
| Claude Fable 5.1anthropic · claude-fable-5-1 | 1,000,000 | 128,000 | $10$0.25 cached | $50 | Vision · Reasoning |
| Gemini 3.7 Flashgoogle · gemini-3.7-flash | 1,048,576 | 65,536 | $0.75$0.075 cached | $3.75 | Vision · Reasoning |
| Gemini Flash Latestgoogle · gemini-flash-latest | 1,048,576 | 65,536 | $0.75$0.075 cached | $3.75 | Vision · Reasoning |
| Claude Opus 5anthropic · claude-opus-5 | 1,000,000 | 128,000 | $5$0.5 cached | $25 | Vision · Reasoning |
| Gemini 3.6 Flashgoogle · gemini-3.6-flash | 1,048,576 | 65,536 | $0.75$0.075 cached | $3.75 | Vision · Reasoning |
| Gemini 3.5 Flash Litegoogle · gemini-3.5-flash-lite | 1,048,576 | 65,536 | $0.3$0.03 cached | $2.5 | Vision · Reasoning |
| Gemini Flash-Lite Latestgoogle · gemini-flash-lite-latest | 1,048,576 | 65,536 | $0.3$0.03 cached | $2.5 | Vision · Reasoning |
GPT-5.6 Solopenai · gpt-5.6-solLong context pricingAbove 272,000 input tokens: $8 input / $30 output per 1M | 1,050,000 | 128,000 | $4$0.4 cached | $20 | Vision · Reasoning |
GPT-5.6 Lunaopenai · gpt-5.6-lunaLong context pricingAbove 272,000 input tokens: $0.4 input / $1.8 output per 1M | 1,050,000 | 128,000 | $0.2$0.02 cached | $1.2 | Vision · Reasoning |
GPT-5.6openai · gpt-5.6Long context pricingAbove 272,000 input tokens: $8 input / $30 output per 1M | 1,050,000 | 128,000 | $4$0.4 cached | $20 | Vision · Reasoning |
GPT-5.6 Terraopenai · gpt-5.6-terraLong context pricingAbove 272,000 input tokens: $4 input / $18 output per 1M | 1,050,000 | 128,000 | $2$0.2 cached | $12 | Vision · Reasoning |
| Claude Sonnet 5anthropic · claude-sonnet-5 | 1,000,000 | 128,000 | $2$0.2 cached | $10 | Vision · Reasoning |
| Claude Fable 5anthropic · claude-fable-5 | 1,000,000 | 128,000 | $10$1 cached | $50 | Vision · Reasoning |
| Claude Opus 4.8anthropic · claude-opus-4-8 | 1,000,000 | 128,000 | $5$0.5 cached | $25 | Vision · Reasoning |
| Gemini 3.5 Flashgoogle · gemini-3.5-flash | 1,048,576 | 65,536 | $1.5$0.15 cached | $9 | Vision · Reasoning |
| Gemini 3.1 Flash Litegoogle · gemini-3.1-flash-lite | 1,048,576 | 65,536 | $0.25$0.025 cached | $1.5 | Vision · Reasoning |
GPT-5.5 Proopenai · gpt-5.5-proLong context pricingAbove 272,000 input tokens: $60 input / $270 output per 1M | 1,050,000 | 128,000 | $30 | $180 | Vision · Reasoning |
GPT-5.5openai · gpt-5.5Long context pricingAbove 272,000 input tokens: $10 input / $45 output per 1M | 1,050,000 | 128,000 | $5$0.5 cached | $30 | Vision · Reasoning |
Deep Research Max Preview (Apr-21-2026)google · deep-research-max-preview-04-2026Long context pricingAbove 200,000 input tokens: $4 input / $18 output per 1M | 131,072 | 65,536 | $2$0.2 cached | $12 | Vision · Reasoning |
Deep Research Preview (Apr-21-2026)google · deep-research-preview-04-2026Long context pricingAbove 200,000 input tokens: $4 input / $18 output per 1M | 131,072 | 65,536 | $2$0.2 cached | $12 | Vision · Reasoning |
| Claude Opus 4.7anthropic · claude-opus-4-7 | 1,000,000 | 128,000 | $5$0.5 cached | $25 | Vision · Reasoning |
| Gemma 4 26B A4B ITgoogle · gemma-4-26b-a4b-it | 262,144 | 32,768 | Unavailable | Unavailable | Vision · Reasoning |
| Gemma 4 31B ITgoogle · gemma-4-31b-it | 262,144 | 32,768 | Unavailable | Unavailable | Vision · Reasoning |
| GPT-5.4 nanoopenai · gpt-5.4-nano | 400,000 | 128,000 | $0.2$0.02 cached | $1.25 | Vision · Reasoning |
| GPT-5.4 miniopenai · gpt-5.4-mini | 400,000 | 128,000 | $0.75$0.075 cached | $4.5 | Vision · Reasoning |
GPT-5.4openai · gpt-5.4Long context pricingAbove 272,000 input tokens: $5 input / $22.5 output per 1M | 1,050,000 | 128,000 | $2.5$0.25 cached | $15 | Vision · Reasoning |
GPT-5.4 Proopenai · gpt-5.4-proLong context pricingAbove 272,000 input tokens: $60 input / $270 output per 1M | 1,050,000 | 128,000 | $30 | $180 | Vision · Reasoning |
| GPT-5.3 Chat (latest)openai · gpt-5.3-chat-latest · deprecated | 128,000 | 16,384 | $1.75$0.175 cached | $14 | Vision |
| Gemini 3.1 Flash Lite Previewgoogle · gemini-3.1-flash-lite-preview · deprecated | 1,048,576 | 65,536 | $0.25$0.025 cached | $1.5 | Vision · Reasoning |
Gemini 3.1 Pro Preview Custom Toolsgoogle · gemini-3.1-pro-preview-customtoolsLong context pricingAbove 200,000 input tokens: $4 input / $18 output per 1M | 1,048,576 | 65,536 | $2$0.2 cached | $12 | Vision · Reasoning |
Gemini 3.1 Pro Previewgoogle · gemini-3.1-pro-previewLong context pricingAbove 200,000 input tokens: $4 input / $18 output per 1M | 1,048,576 | 65,536 | $2$0.2 cached | $12 | Vision · Reasoning |
| Claude Sonnet 4.6anthropic · claude-sonnet-4-6 | 1,000,000 | 128,000 | $3$0.3 cached | $15 | Vision · Reasoning |
| GPT-5.3 Codex Sparkopenai · gpt-5.3-codex-spark | 128,000 | 32,000 | $1.75$0.175 cached | $14 | Vision · Reasoning |
| GPT-5.3 Codexopenai · gpt-5.3-codex | 400,000 | 128,000 | $1.75$0.175 cached | $14 | Vision · Reasoning |
| Claude Opus 4.6anthropic · claude-opus-4-6 | 1,000,000 | 128,000 | $5$0.5 cached | $25 | Vision · Reasoning |
| Gemini 3 Flash Previewgoogle · gemini-3-flash-preview | 1,048,576 | 65,536 | $0.5$0.05 cached | $3 | Vision · Reasoning |
| GPT-5.2 Proopenai · gpt-5.2-pro | 400,000 | 128,000 | $21 | $168 | Vision · Reasoning |
| GPT-5.2openai · gpt-5.2 | 400,000 | 128,000 | $1.75$0.175 cached | $14 | Vision · Reasoning |
| GPT-5.2 Chatopenai · gpt-5.2-chat-latest · deprecated | 128,000 | 16,384 | $1.75$0.175 cached | $14 | Vision · Reasoning |
| Claude Opus 4.5 (latest)anthropic · claude-opus-4-5 | 200,000 | 64,000 | $5$0.5 cached | $25 | Vision · Reasoning |
| Claude Opus 4.5anthropic · claude-opus-4-5-20251101 | 200,000 | 64,000 | $5$0.5 cached | $25 | Vision · Reasoning |
| GPT-5.1openai · gpt-5.1 | 400,000 | 128,000 | $1.25$0.125 cached | $10 | Vision · Reasoning |
| Claude Haiku 4.5anthropic · claude-haiku-4-5-20251001 | 200,000 | 64,000 | $1$0.1 cached | $5 | Vision · Reasoning |
| Claude Haiku 4.5 (latest)anthropic · claude-haiku-4-5 | 200,000 | 64,000 | $1$0.1 cached | $5 | Vision · Reasoning |
Gemini 2.5 Computer Use Preview 10-2025google · gemini-2.5-computer-use-preview-10-2025Long context pricingAbove 200,000 input tokens: $2.5 input / $15 output per 1M | 131,072 | 65,536 | $1.25 | $10 | Vision · Reasoning |
| GPT-5 Proopenai · gpt-5-pro | 400,000 | 272,000 | $15 | $120 | Vision · Reasoning |
| Claude Sonnet 4.5anthropic · claude-sonnet-4-5-20250929 | 1,000,000 | 64,000 | $3$0.3 cached | $15 | Vision · Reasoning |
| Claude Sonnet 4.5 (latest)anthropic · claude-sonnet-4-5 | 1,000,000 | 64,000 | $3$0.3 cached | $15 | Vision · Reasoning |
| GPT-5 Nanoopenai · gpt-5-nano | 400,000 | 128,000 | $0.05$0.005 cached | $0.4 | Vision · Reasoning |
| GPT-5 Miniopenai · gpt-5-mini | 400,000 | 128,000 | $0.25$0.025 cached | $2 | Vision · Reasoning |
| GPT-5openai · gpt-5 | 400,000 | 128,000 | $1.25$0.125 cached | $10 | Vision · Reasoning |
| Gemini 2.5 Flash-Litegoogle · gemini-2.5-flash-lite | 1,048,576 | 65,536 | $0.1$0.01 cached | $0.4 | Vision · Reasoning |
Gemini 2.5 Progoogle · gemini-2.5-proLong context pricingAbove 200,000 input tokens: $2.5 input / $15 output per 1M | 1,048,576 | 65,536 | $1.25$0.125 cached | $10 | Vision · Reasoning |
| Gemini 2.5 Flashgoogle · gemini-2.5-flash | 1,048,576 | 65,536 | $0.3$0.03 cached | $2.5 | Vision · Reasoning |
| o3-proopenai · o3-pro | 200,000 | 100,000 | $20 | $80 | Vision · Reasoning |
| o4-miniopenai · o4-mini · deprecated | 200,000 | 100,000 | $1.1$0.275 cached | $4.4 | Vision · Reasoning |
| o3openai · o3 | 200,000 | 100,000 | $2$0.5 cached | $8 | Vision · Reasoning |
| GPT-4.1 nanoopenai · gpt-4.1-nano · deprecated | 1,047,576 | 32,768 | $0.1$0.025 cached | $0.4 | Vision |
| GPT-4.1 miniopenai · gpt-4.1-mini | 1,047,576 | 32,768 | $0.4$0.1 cached | $1.6 | Vision |
| GPT-4.1openai · gpt-4.1 | 1,047,576 | 32,768 | $2$0.5 cached | $8 | Vision |
| o1-proopenai · o1-pro · deprecated | 200,000 | 100,000 | $150 | $600 | Vision · Reasoning |
| o3-miniopenai · o3-mini · deprecated | 200,000 | 100,000 | $1.1$0.55 cached | $4.4 | Reasoning |
| o1openai · o1 · deprecated | 200,000 | 100,000 | $15$7.5 cached | $60 | Vision · Reasoning |
| GPT-4o (2024-11-20)openai · gpt-4o-2024-11-20 | 128,000 | 16,384 | $2.5$1.25 cached | $10 | Vision |
| GPT-4o (2024-08-06)openai · gpt-4o-2024-08-06 | 128,000 | 16,384 | $2.5$1.25 cached | $10 | Vision |
| GPT-4o miniopenai · gpt-4o-mini | 128,000 | 16,384 | $0.15$0.075 cached | $0.6 | Vision |
| GPT-4o (2024-05-13)openai · gpt-4o-2024-05-13 · deprecated | 128,000 | 4,096 | $5 | $15 | Vision |
| GPT-4oopenai · gpt-4o | 128,000 | 16,384 | $2.5$1.25 cached | $10 | Vision |
| GPT-4 Turboopenai · gpt-4-turbo · deprecated | 128,000 | 4,096 | $10 | $30 | Vision |
| GPT-4openai · gpt-4 · deprecated | 8,192 | 8,192 | $30 | $60 | Text |
| GPT-3.5-turboopenai · gpt-3.5-turbo · deprecated | 16,385 | 4,096 | $0.5$0 cached | $1.5 | Text |
Refreshing model data… · Models.dev · 2026-09-12Community maintained metadata and standard USD pricing. Provider terms and availability can differ.
A prompt is more than its text.
Messages, tool definitions, files, and images add input tokens. A local tokenizer counts the text you entered. A provider count measures the request you actually send.
Input and output have different prices.
The counter estimates input only. Generated output, reasoning tokens, tools, caching, batch discounts, and long context pricing can change a real bill.
Images need their own rules.
There is no universal tokens-per-image or tokens-per-second formula. The model, image detail, resolution, and modality determine the count.
Explore image tokens →Sources and accuracy
The local OpenAI counter is an encoding estimate, not a billing measurement. New or unsupported tokenizers display no invented count.
OpenAI token counting · Claude token counting · Gemini countTokens
The directory refreshes on a visit after its 24 hour cache expires. A saved copy remains available if the feed fails. A refresh timestamp indicates retrieval, not independent verification of every model field. Model availability depends on the provider account.