Skip to main content
TokenCost logoTokenCost

How many tokens per dollar?

At the median API price (October 2026), $1 buys about 1,052,632 input tokens or 250,000 output tokens. The cheapest output is Qwen3.7 Flash at 7,692,308 tokens per dollar. The most expensive is o1 Pro at 1,667.

The formula

Tokens per dollar = 1,000,000 ÷ price per million tokens. Work it out twice, once for input and once for output, because output usually costs several times more. Every figure below comes from the same price catalogue as the pricing table, so it updates when a price changes. 163 paid models are included; models with no paid API are left out.

Tokens are not words. 1 million tokens is about 750,000 English words. To convert, use the tokens to words converter.

Anthropic tokens per dollar

ModelInput tokens per $1Output tokens per $1Output tokens for $20
Claude Haiku 5.510,000,0002,000,00040,000,000
Claude Haiku 4.51,000,000200,0004,000,000
Claude Sonnet 5.5500,000100,0002,000,000
Claude Sonnet 5500,000100,0002,000,000
Claude Sonnet 4.6333,33366,6671,333,333
Claude Sonnet 4.5333,33366,6671,333,333
Claude Opus 5.5250,00050,0001,000,000
Claude Opus 5200,00040,000800,000
Claude Opus 4.8200,00040,000800,000
Claude Opus 4.7200,00040,000800,000
Claude Opus 4.6200,00040,000800,000
Claude Opus 4.5200,00040,000800,000
Claude Fable 5.1100,00020,000400,000
Claude Fable 5100,00020,000400,000

OpenAI tokens per dollar

ModelInput tokens per $1Output tokens per $1Output tokens for $20
GPT-OSS 120B †25,641,0265,263,158105,263,158
GPT-OSS 20B †13,333,3333,333,33366,666,667
GPT-OSS 20B (Bedrock) †14,285,7143,333,33366,666,667
GPT-5 Nano20,000,0002,500,00050,000,000
GPT-4.1 Nano10,000,0002,500,00050,000,000
GPT-6 Luna10,000,0002,000,00040,000,000
GPT-OSS 120B (Bedrock) †6,666,6671,666,66733,333,333
GPT-4o Mini6,666,6671,666,66733,333,333
GPT-5.6 Luna5,000,000833,33316,666,667
GPT-5.4 Nano5,000,000800,00016,000,000
GPT-4.1 Mini2,500,000625,00012,500,000
GPT-5 Mini4,000,000500,00010,000,000
o4 Mini909,091227,2734,545,455
o3 Mini909,091227,2734,545,455
GPT-5.4 Mini1,333,333222,2224,444,444
GPT-6.1 Terra1,000,000200,0004,000,000
GPT-4.1500,000125,0002,500,000
o3500,000125,0002,500,000
o4 Mini Deep Research †500,000125,0002,500,000
GPT-6.1 Sol500,000100,0002,000,000
GPT-6 Sol500,000100,0002,000,000
GPT-5.1800,000100,0002,000,000
GPT-5800,000100,0002,000,000
GPT-4o400,000100,0002,000,000
GPT-5.6 Terra500,00083,3331,666,667
GPT-5.2571,42971,4291,428,571
GPT-5.3 Codex571,42971,4291,428,571
GPT-5.4400,00066,6671,333,333
GPT-5.6 Sol250,00050,0001,000,000
GPT-5.5200,00033,333666,667
GPT Image 2.5200,00033,333666,667
GPT Image 2200,00033,333666,667
o3 Deep Research †100,00025,000500,000
o166,66716,667333,333
GPT-5.6 Cyber80,00013,333266,667
o3-pro50,00012,500250,000
GPT-5.5 Pro33,3335,556111,111
GPT-5.4 Pro33,3335,556111,111
o1 Pro6,6671,66733,333

Google tokens per dollar

ModelInput tokens per $1Output tokens per $1Output tokens for $20
Gemma 4 26B A4B †16,666,6673,333,33366,666,667
Gemma 4 31B †8,333,3332,777,77855,555,556
Gemini 2.5 Flash-Lite10,000,0002,500,00050,000,000
Gemini 3.1 Flash-Lite4,000,000666,66713,333,333
Gemini 3.5 Flash-Lite3,333,333400,0008,000,000
Gemini 2.5 Flash3,333,333400,0008,000,000
Gemini 3 Flash2,000,000333,3336,666,667
Gemini 3.8 Flash1,333,333266,6675,333,333
Gemini 3.6 Flash1,333,333266,6675,333,333
Gemini 3.7 Flash1,333,333266,6675,333,333
Gemini 3.1 Flash Live1,333,333222,2224,444,444
Gemini Omni Flash666,667111,1112,222,222
Gemini Omni Flash (Preview)666,667111,1112,222,222
Gemini 3.5 Flash666,667111,1112,222,222
Gemini 4 Argon500,000100,0002,000,000
Gemini 2.5 Pro800,000100,0002,000,000
Gemini 3.1 Pro500,00083,3331,666,667
Gemini 3.1 Flash TTS1,000,00050,0001,000,000
Nano Banana 2.1666,66733,333666,667
Gemini 3.1 Flash Lite Image4,000,00033,333666,667
Gemini 3.1 Flash Image2,000,00016,667333,333

DeepSeek tokens per dollar

ModelInput tokens per $1Output tokens per $1Output tokens for $20
DeepSeek V4.1 Flash6,666,6671,666,66733,333,333
DeepSeek V4-Pro1,515,152505,05110,101,010

xAI tokens per dollar

ModelInput tokens per $1Output tokens per $1Output tokens for $20
Grok Build 0.11,000,000500,00010,000,000
Grok 4.3800,000400,0008,000,000
Grok 4.20800,000400,0008,000,000
Grok 4.7500,000166,6673,333,333
Grok 4.6500,000166,6673,333,333
Grok 4.5500,000166,6673,333,333

All models, most output tokens per dollar first

ModelInput tokens per $1Output tokens per $1Output tokens for $20
Qwen3.7 Flash33,333,3337,692,308153,846,154
Ministral 3 8B6,666,6676,666,667133,333,333
Qwen3.5-9B10,000,0006,666,667133,333,333
Mercury 2.5 (Preview)25,000,0006,666,667133,333,333
Llama 3.3 70B5,555,5565,555,556111,111,111
Ling-3.0-flash16,666,6675,555,556111,111,111
GPT-OSS 120B †25,641,0265,263,158105,263,158
Muse Spark 1.3 Contributor10,000,0005,000,000100,000,000
Muse Spark 1.2 Contributor †10,000,0005,000,000100,000,000
Ministral 3 14B5,000,0005,000,000100,000,000
Laguna S 2.110,000,0005,000,000100,000,000
MiMo-V2.6-Flash7,142,8573,571,42971,428,571
GPT-OSS 20B †13,333,3333,333,33366,666,667
GPT-OSS 20B (Bedrock) †14,285,7143,333,33366,666,667
Gemma 4 26B A4B †16,666,6673,333,33366,666,667
Llama 4 Scout12,500,0003,333,33366,666,667
Gemma 4 31B †8,333,3332,777,77855,555,556
GPT-5 Nano20,000,0002,500,00050,000,000
GPT-4.1 Nano10,000,0002,500,00050,000,000
Gemini 2.5 Flash-Lite10,000,0002,500,00050,000,000
GLM-4.7-FlashX14,285,7142,500,00050,000,000
Qwen3.8 Flash6,666,6672,127,66042,553,191
GPT-6 Luna10,000,0002,000,00040,000,000
Claude Haiku 5.510,000,0002,000,00040,000,000
GLM-5.3-Flash6,666,6672,000,00040,000,000
Hy37,575,7581,893,93937,878,788
Hunyuan HY3 Preview †5,617,9781,680,67233,613,445
GPT-OSS 120B (Bedrock) †6,666,6671,666,66733,333,333
GPT-4o Mini6,666,6671,666,66733,333,333
Mistral Small 46,666,6671,666,66733,333,333
DeepSeek V4.1 Flash6,666,6671,666,66733,333,333
KAT-Coder-Air v2.56,666,6671,666,66733,333,333
MiniMax M3.1 Flash6,666,6671,666,66733,333,333
Mercury 24,000,0001,333,33326,666,667
Qwen3 Coder Next8,333,3331,250,00025,000,000
Nemotron 3 Super 120B3,333,3331,250,00025,000,000
Llama 4 Maverick3,703,7041,176,47123,529,412
MiMo-V2.6-Pro2,298,8511,149,42522,988,506
MiMo-V2.5-Pro2,298,8511,149,42522,988,506
Codestral3,333,3331,111,11122,222,222
Step 3.7 Flash5,000,000869,56517,391,304
GPT-5.6 Luna5,000,000833,33316,666,667
Qwen3-Next-80B-A3B-Thinking6,666,667833,33316,666,667
MiniMax M33,333,333833,33316,666,667
MiniMax M2.73,333,333833,33316,666,667
MiniMax M2.53,333,333833,33316,666,667
GPT-5.4 Nano5,000,000800,00016,000,000
Gemini 3.1 Flash-Lite4,000,000666,66713,333,333
Muse Glimmer 30B2,857,143666,66713,333,333
Mistral Large 32,000,000666,66713,333,333
GPT-4.1 Mini2,500,000625,00012,500,000
Qwen3.7 Plus2,500,000625,00012,500,000
DeepSeek V4-Pro1,515,152505,05110,101,010
GPT-5 Mini4,000,000500,00010,000,000
Grok Build 0.11,000,000500,00010,000,000
Qwen3.5-Omni Flash2,500,000454,5459,090,909
GLM-4.71,666,667454,5459,090,909
Doubao Seed 2.1 Turbo2,272,727452,4899,049,774
Qwen3.5-27B3,333,333416,6678,333,333
QwQ-Plus1,250,000416,6678,333,333
Gemini 3.5 Flash-Lite3,333,333400,0008,000,000
Gemini 2.5 Flash3,333,333400,0008,000,000
Grok 4.3800,000400,0008,000,000
Grok 4.20800,000400,0008,000,000
Nova 2.0 Lite3,333,333400,0008,000,000
Nemotron 3 Ultra 550B2,000,000400,0008,000,000
Nex-N2-Pro2,000,000400,0008,000,000
Hy4 preview1,199,041399,8407,996,801
LongCat-2.01,333,333338,9836,779,661
KAT-Coder-Pro v2.51,351,351337,8386,756,757
Gemini 3 Flash2,000,000333,3336,666,667
Qwen3.6-Plus2,000,000333,3336,666,667
Kimi K2.51,666,667333,3336,666,667
MiMo-V2-Pro1,000,000333,3336,666,667
GLM-51,000,000312,5006,250,000
Qwen3.5-397B-A17B1,666,667277,7785,555,556
Gemini 3.8 Flash1,333,333266,6675,333,333
Gemini 3.6 Flash1,333,333266,6675,333,333
Gemini 3.7 Flash1,333,333266,6675,333,333
Kimi K2.7 Code1,052,632250,0005,000,000
Kimi K2.61,052,632250,0005,000,000
GLM-5 Turbo833,333250,0005,000,000
Mistral Large 4735,294239,2344,784,689
Muse Spark 1.1800,000235,2944,705,882
Muse Spark 1.3800,000235,2944,705,882
Muse Spark 1.2 †800,000235,2944,705,882
Muse Spark Open800,000235,2944,705,882
o4 Mini909,091227,2734,545,455
o3 Mini909,091227,2734,545,455
GLM-5.3714,286227,2734,545,455
GLM-5.2714,286227,2734,545,455
GLM-5.1714,286227,2734,545,455
Doubao Seed 2.1 Pro1,136,364226,7574,535,147
GPT-5.4 Mini1,333,333222,2224,444,444
Gemini 3.1 Flash Live1,333,333222,2224,444,444
Qwen3.5-Omni Plus2,500,000208,3334,166,667
GPT-6.1 Terra1,000,000200,0004,000,000
Claude Haiku 4.51,000,000200,0004,000,000
Grok 4.7500,000166,6673,333,333
Grok 4.6500,000166,6673,333,333
Grok 4.5500,000166,6673,333,333
Qwen3.8 Max500,000166,6673,333,333
Fugu Max500,000166,6673,333,333
Mistral Medium 3.5666,667133,3332,666,667
Qwen3.7 Max400,000133,3332,666,667
Qwen3.6-Max-Preview769,231128,2052,564,103
GPT-4.1500,000125,0002,500,000
o3500,000125,0002,500,000
o4 Mini Deep Research †500,000125,0002,500,000
GLM-5.3 Prime357,143113,6362,272,727
Gemini Omni Flash666,667111,1112,222,222
Gemini Omni Flash (Preview)666,667111,1112,222,222
Gemini 3.5 Flash666,667111,1112,222,222
GPT-6.1 Sol500,000100,0002,000,000
GPT-6 Sol500,000100,0002,000,000
GPT-5.1800,000100,0002,000,000
GPT-5800,000100,0002,000,000
GPT-4o400,000100,0002,000,000
Claude Sonnet 5.5500,000100,0002,000,000
Claude Sonnet 5500,000100,0002,000,000
Gemini 4 Argon500,000100,0002,000,000
Gemini 2.5 Pro800,000100,0002,000,000
Nova 2.0 Pro Reasoning †800,000100,0002,000,000
Command A+400,000100,0002,000,000
Command A400,000100,0002,000,000
Kimi K3.1 Pro500,000100,0002,000,000
GPT-5.6 Terra500,00083,3331,666,667
Gemini 3.1 Pro500,00083,3331,666,667
Qwen3.8 Max Prime250,00083,3331,666,667
GPT-5.2571,42971,4291,428,571
GPT-5.3 Codex571,42971,4291,428,571
GPT-5.4400,00066,6671,333,333
Claude Sonnet 4.6333,33366,6671,333,333
Claude Sonnet 4.5333,33366,6671,333,333
Sonar Pro333,33366,6671,333,333
Kimi K3333,33366,6671,333,333
GPT-5.6 Sol250,00050,0001,000,000
Claude Opus 5.5250,00050,0001,000,000
Gemini 3.1 Flash TTS1,000,00050,0001,000,000
Muse Wonder250,00050,0001,000,000
Kimi K3.1 Pro Highspeed250,00050,0001,000,000
Claude Opus 5200,00040,000800,000
Claude Opus 4.8200,00040,000800,000
Claude Opus 4.7200,00040,000800,000
Claude Opus 4.6200,00040,000800,000
Claude Opus 4.5200,00040,000800,000
GPT-5.5200,00033,333666,667
GPT Image 2.5200,00033,333666,667
GPT Image 2200,00033,333666,667
Nano Banana 2.1666,66733,333666,667
Gemini 3.1 Flash Lite Image4,000,00033,333666,667
Fugu Ultra v2200,00033,333666,667
MAI-Image-2200,00030,303606,061
o3 Deep Research †100,00025,000500,000
Claude Fable 5.1100,00020,000400,000
Claude Fable 5100,00020,000400,000
o166,66716,667333,333
Gemini 3.1 Flash Image2,000,00016,667333,333
GPT-5.6 Cyber80,00013,333266,667
o3-pro50,00012,500250,000
GPT-5.5 Pro33,3335,556111,111
GPT-5.4 Pro33,3335,556111,111
o1 Pro6,6671,66733,333

† The provider publishes no rate for this model, so its price comes from a weaker source. List prices only. Prompt caching and batch discounts buy more tokens per dollar on providers that offer them. The cost calculator models both.

Related Tools

API Pricing

Price per 1M tokens for every model.

Cheapest LLMs

Models ranked by average price.

Tokens to Words

Convert tokens to words and pages.

Cost Calculator

Estimate monthly API spend.

Frequently Asked Questions

How many tokens do you get for $1?
It depends on the model. At the median price among 163 paid models, $1 buys about 1,052,632 input tokens or 250,000 output tokens. The range runs from 1,667 output tokens on o1 Pro to 7,692,308 on Qwen3.7 Flash.
How do I work out tokens per dollar?
Divide 1,000,000 by the price per million tokens. A model at $2 per 1M tokens gives 500,000 tokens per dollar. Do it separately for input and output, because they are priced differently.
How much are 1,000 tokens worth?
Divide the per-million price by 1,000. At the median paid price that is $0.00095 for 1,000 input tokens and $0.00400 for 1,000 output tokens.
Is a Claude Pro or ChatGPT subscription priced per token?
No. Consumer subscriptions are flat monthly fees with usage limits, not per-token billing. The figures on this page are for the pay-as-you-go APIs.