Official

anthropic / claude-3.5-sonnet

Anthropic's most intelligent language model to date, with a 200K token context window and image understanding (claude-3-5-sonnet-20241022)

Warm

Public
324.8K runs
Priced per token
Commercial use
License

Run with an API

Playground API Examples README

Prediction

anthropic/claude-3.5-sonnet

Official model

anthropic/claude-3.5-sonnet

8bwz4aszy5rm80cmzbttpakyfc

Status

Succeeded

Source

Web

Total duration

6.0s

Created

about 2 months ago

Input

prompt: Give me a recipe for pancakes that could feed all of California.
max_tokens: 8192
system_prompt
max_image_resolution: 0.5

{
  "prompt": "Give me a recipe for pancakes that could feed all of California.",
  "max_tokens": 8192,
  "system_prompt": "",
  "max_image_resolution": 0.5
}

Install Replicate’s Node.js client library:

npm install replicate

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import and set up the client:

import Replicate from "replicate";

const replicate = new Replicate({
  auth: process.env.REPLICATE_API_TOKEN,
});

Run anthropic/claude-3.5-sonnet using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const input = {
  prompt: "Give me a recipe for pancakes that could feed all of California.",
  max_tokens: 8192,
  system_prompt: "",
  max_image_resolution: 0.5
};

for await (const event of replicate.stream("anthropic/claude-3.5-sonnet", { input })) {
  process.stdout.write(event.toString());
};

To learn more, take a look at the guide on getting started with Node.js.

Install Replicate’s Python client library:

pip install replicate

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import the client:

import replicate

Run anthropic/claude-3.5-sonnet using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

# The anthropic/claude-3.5-sonnet model can stream output as it's running.
for event in replicate.stream(
    "anthropic/claude-3.5-sonnet",
    input={
        "prompt": "Give me a recipe for pancakes that could feed all of California.",
        "max_tokens": 8192,
        "system_prompt": "",
        "max_image_resolution": 0.5
    },
):
    print(str(event), end="")

To learn more, take a look at the guide on getting started with Python.

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Run anthropic/claude-3.5-sonnet using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

curl -s -X POST \
  -H "Authorization: Bearer $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -H "Prefer: wait" \
  -d $'{
    "input": {
      "prompt": "Give me a recipe for pancakes that could feed all of California.",
      "max_tokens": 8192,
      "system_prompt": "",
      "max_image_resolution": 0.5
    }
  }' \
  https://api.replicate.com/v1/models/anthropic/claude-3.5-sonnet/predictions

To learn more, take a look at Replicate’s HTTP API reference docs.

Output

Here's a humorous yet enormous pancake recipe scaled up for California's population (approximately 40 million people): "California-Sized Pancake Recipe" Ingredients: * 20 million cups all-purpose flour (about 2,500 tons) * 40 million eggs (need about 3.3 million chickens working overtime) * 15 million cups milk (about 240,000 cows' daily production) * 10 million tablespoons baking powder * 5 million teaspoons salt * 5 million cups sugar * 5 million cups melted butter You'll need: * The entire Central Valley as a mixing bowl * Several helicopters with giant whisks * A tectonic plate-sized griddle * The Pacific Ocean converted to maple syrup Instructions: 1. Mix dry ingredients in Death Valley 2. Combine wet ingredients in Lake Tahoe 3. Pour batter using the California Aqueduct 4. Cook on the San Andreas Fault line for even heat distribution 5. Flip using coordinated effort from entire US Air Force Warning: May cause minor earthquakes when flipping. Not responsible for syrup-related flooding. Serves: 40 million Californians, plus leftovers for Nevada Note: This is obviously a silly recipe - it would be logistically impossible to make and cook a pancake this large! 😄

{
  "completed_at": "2025-02-12T16:19:30.270847Z",
  "created_at": "2025-02-12T16:19:24.273000Z",
  "data_removed": false,
  "error": null,
  "id": "8bwz4aszy5rm80cmzbttpakyfc",
  "input": {
    "prompt": "Give me a recipe for pancakes that could feed all of California.",
    "max_tokens": 8192,
    "system_prompt": "",
    "max_image_resolution": 0.5
  },
  "logs": "Running prediction...",
  "metrics": {
    "input_token_count": 21,
    "tokens_per_second": 54.00962598247224,
    "output_token_count": 323,
    "predict_time": 5.982320888,
    "total_time": 5.997847,
    "time_to_first_token": 0.017431468999999998
  },
  "output": [
    "Here's a hum",
    "orous yet enormous",
    " pancake recipe scale",
    "d up for California's population (",
    "approximately 40 million people",
    "):\n\n\"California-",
    "Sized Pancake",
    " Recipe\"\n\nIngredients",
    ":\n* 20",
    " million cups all-purpose flour (",
    "about 2,500",
    " tons)\n* 40",
    " million eggs (nee",
    "d about 3",
    ".3 million chickens working",
    " overtime)\n* 15",
    " million cups milk (about 240",
    ",000 cows'",
    " daily production)\n* 10",
    " million tablespoons b",
    "aking powder\n* 5 million",
    " teaspoons salt\n*",
    " 5 million cups sugar",
    "\n* 5 million cups mel",
    "ted butter\n\nYou'll nee",
    "d:\n* The entire Central",
    " Valley as a mixing bowl\n* Several",
    " helicopters with giant whisks\n* A",
    " tectonic plate-",
    "sized griddle\n*",
    " The Pacific Ocean converted to",
    " maple syrup\n\nInstructions",
    ":\n1. Mix dry",
    " ingredients in Death Valley",
    "\n2. Combine wet ingredients in Lake",
    " Tahoe\n3. Pour",
    " batter using the California",
    " Aqueduct\n4. Cook",
    " on the San Andreas Fault line",
    " for even heat distribution\n5.",
    " Flip using coordinated effort from entire",
    " US Air Force\n\nWarning",
    ": May cause minor earthquakes when",
    " flipping. Not responsible for sy",
    "rup-related flooding.\n\nServes",
    ": 40 million Californ",
    "ians, plus leftovers for Nevada",
    "\n\nNote: This is obviously",
    " a silly recipe - it woul",
    "d be logistically impossible to make an",
    "d cook a pancake this large!",
    " 😄"
  ],
  "started_at": "2025-02-12T16:19:24.288526Z",
  "status": "succeeded",
  "urls": {
    "stream": "https://stream-b.svc.ric1.c.replicate.net/v1/streams/fj6cmzuqif63sszp4whb5gdzmptgjvqz3r6wyge2gqo7wl4gxzja",
    "get": "https://api.replicate.com/v1/predictions/8bwz4aszy5rm80cmzbttpakyfc",
    "cancel": "https://api.replicate.com/v1/predictions/8bwz4aszy5rm80cmzbttpakyfc/cancel"
  },
  "version": "hidden"
}

Generated in

6.0 seconds

Input tokens

Output tokens

320

Tokens per second

54.01 tokens / second

Time to first token

17 milliseconds

Tweak it Share

Prediction

anthropic/claude-3.5-sonnet

Official model

anthropic/claude-3.5-sonnet

p5nes79jvhrma0cmzbv8c3dx38

Status

Succeeded

Source

Web

Total duration

0.6s

Created

about 2 months ago

Input

prompt: (d/dx) * (x * cos(x))
max_tokens: 8192
system_prompt: You are a calculator. You only output the answer to my question, without backticks or explanatory text, as I intend to use your output programmatically.
max_image_resolution: 0.5

{
  "prompt": "(d/dx) * (x * cos(x))",
  "max_tokens": 8192,
  "system_prompt": "You are a calculator. You only output the answer to my question, without backticks or explanatory text, as I intend to use your output programmatically.",
  "max_image_resolution": 0.5
}

Install Replicate’s Node.js client library:

npm install replicate

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import and set up the client:

import Replicate from "replicate";

const replicate = new Replicate({
  auth: process.env.REPLICATE_API_TOKEN,
});

Run anthropic/claude-3.5-sonnet using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const input = {
  prompt: "(d/dx) * (x * cos(x))",
  max_tokens: 8192,
  system_prompt: "You are a calculator. You only output the answer to my question, without backticks or explanatory text, as I intend to use your output programmatically.",
  max_image_resolution: 0.5
};

for await (const event of replicate.stream("anthropic/claude-3.5-sonnet", { input })) {
  process.stdout.write(event.toString());
};

To learn more, take a look at the guide on getting started with Node.js.

Install Replicate’s Python client library:

pip install replicate

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import the client:

import replicate

Run anthropic/claude-3.5-sonnet using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

# The anthropic/claude-3.5-sonnet model can stream output as it's running.
for event in replicate.stream(
    "anthropic/claude-3.5-sonnet",
    input={
        "prompt": "(d/dx) * (x * cos(x))",
        "max_tokens": 8192,
        "system_prompt": "You are a calculator. You only output the answer to my question, without backticks or explanatory text, as I intend to use your output programmatically.",
        "max_image_resolution": 0.5
    },
):
    print(str(event), end="")

To learn more, take a look at the guide on getting started with Python.

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Run anthropic/claude-3.5-sonnet using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

curl -s -X POST \
  -H "Authorization: Bearer $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -H "Prefer: wait" \
  -d $'{
    "input": {
      "prompt": "(d/dx) * (x * cos(x))",
      "max_tokens": 8192,
      "system_prompt": "You are a calculator. You only output the answer to my question, without backticks or explanatory text, as I intend to use your output programmatically.",
      "max_image_resolution": 0.5
    }
  }' \
  https://api.replicate.com/v1/models/anthropic/claude-3.5-sonnet/predictions

To learn more, take a look at Replicate’s HTTP API reference docs.

Output

cos(x) - x*sin(x)

Generated in

0.6 seconds

Input tokens

Output tokens

Tokens per second

24.62 tokens / second

Time to first token

8 milliseconds

Tweak it Share

Prediction

anthropic/claude-3.5-sonnet

Official model

anthropic/claude-3.5-sonnet

vhjdxs480hrme0cmzbvbk4xc0r

Status

Succeeded

Source

Web

Total duration

16.9s

Created

about 2 months ago

Input

image
prompt: Implement this website in HTML as accurately as possible.
max_tokens: 8192
system_prompt: Only respond with the requested HTML, no backticks, no explanations, as I intend to use your output in a programmatic context.
max_image_resolution: 0.5

{
  "image": "https://replicate.delivery/pbxt/MTrw3BwfRvOKtZMROPusfnTfwh2j9Ew0aruSY8u4VbWuctp8/image.png",
  "prompt": "Implement this website in HTML as accurately as possible.",
  "max_tokens": 8192,
  "system_prompt": "Only respond with the requested HTML, no backticks, no explanations, as I intend to use your output in a programmatic context.",
  "max_image_resolution": 0.5
}

Install Replicate’s Node.js client library:

npm install replicate

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import and set up the client:

import Replicate from "replicate";

const replicate = new Replicate({
  auth: process.env.REPLICATE_API_TOKEN,
});

Run anthropic/claude-3.5-sonnet using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const input = {
  image: "https://replicate.delivery/pbxt/MTrw3BwfRvOKtZMROPusfnTfwh2j9Ew0aruSY8u4VbWuctp8/image.png",
  prompt: "Implement this website in HTML as accurately as possible.",
  max_tokens: 8192,
  system_prompt: "Only respond with the requested HTML, no backticks, no explanations, as I intend to use your output in a programmatic context.",
  max_image_resolution: 0.5
};

for await (const event of replicate.stream("anthropic/claude-3.5-sonnet", { input })) {
  process.stdout.write(event.toString());
};

To learn more, take a look at the guide on getting started with Node.js.

Install Replicate’s Python client library:

pip install replicate

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import the client:

import replicate

Run anthropic/claude-3.5-sonnet using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

# The anthropic/claude-3.5-sonnet model can stream output as it's running.
for event in replicate.stream(
    "anthropic/claude-3.5-sonnet",
    input={
        "image": "https://replicate.delivery/pbxt/MTrw3BwfRvOKtZMROPusfnTfwh2j9Ew0aruSY8u4VbWuctp8/image.png",
        "prompt": "Implement this website in HTML as accurately as possible.",
        "max_tokens": 8192,
        "system_prompt": "Only respond with the requested HTML, no backticks, no explanations, as I intend to use your output in a programmatic context.",
        "max_image_resolution": 0.5
    },
):
    print(str(event), end="")

To learn more, take a look at the guide on getting started with Python.

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Run anthropic/claude-3.5-sonnet using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

curl -s -X POST \
  -H "Authorization: Bearer $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -H "Prefer: wait" \
  -d $'{
    "input": {
      "image": "https://replicate.delivery/pbxt/MTrw3BwfRvOKtZMROPusfnTfwh2j9Ew0aruSY8u4VbWuctp8/image.png",
      "prompt": "Implement this website in HTML as accurately as possible.",
      "max_tokens": 8192,
      "system_prompt": "Only respond with the requested HTML, no backticks, no explanations, as I intend to use your output in a programmatic context.",
      "max_image_resolution": 0.5
    }
  }' \
  https://api.replicate.com/v1/models/anthropic/claude-3.5-sonnet/predictions

To learn more, take a look at Replicate’s HTTP API reference docs.

Output

<html> <head> <title>Welcome to My CYBER-SPACE</title> <style> body { background: url('space-bg.jpg'); color: white; font-family: Arial, sans-serif; } .header { text-align: center; } .logo { font-family: "Comic Sans MS", cursive; color: #00ffff; } .main-content { width: 800px; margin: 0 auto; background: rgba(0, 0, 0, 0.7); padding: 20px; border: 2px solid #4169e1; border-radius: 10px; } .button { background: #808080; color: white; padding: 5px 15px; margin: 5px; border: 1px solid #666; border-radius: 3px; } .guestbook { border: 1px solid #4169e1; padding: 10px; margin: 10px 0; } .navigation { display: flex; justify-content: center; gap: 10px; margin: 20px 0; } </style> </head> <body> <div class="header"> <div class="logo">Rono!</div> <h1>Welcome to</h1> <h2>My</h2> <h2 style="color: red;">CYBER-SPACE</h2> </div> <div class="main-content"> <div class="navigation"> <button class="button">Rrpcre Sapes</button> <button class="button">Art Lygpduk</button> <button class="button">Sed las</button> </div> <div class="content-area"> <div class="guestbook"> <h3>GUESTBOOK</h3> <form> <textarea rows="4" cols="50"></textarea> </form> </div> </div> <div class="navigation"> <button class="button">Canay? Syargh</button> <button class="button">Apor tov daybe</button> </div> </div> </body> </html>

{
  "completed_at": "2025-02-12T16:21:05.170512Z",
  "created_at": "2025-02-12T16:20:48.260000Z",
  "data_removed": false,
  "error": null,
  "id": "vhjdxs480hrme0cmzbvbk4xc0r",
  "input": {
    "image": "https://replicate.delivery/pbxt/MTrw3BwfRvOKtZMROPusfnTfwh2j9Ew0aruSY8u4VbWuctp8/image.png",
    "prompt": "Implement this website in HTML as accurately as possible.",
    "max_tokens": 8192,
    "system_prompt": "Only respond with the requested HTML, no backticks, no explanations, as I intend to use your output in a programmatic context.",
    "max_image_resolution": 0.5
  },
  "logs": "Images are equivalent to 666 input tokens\nRunning prediction...",
  "metrics": {
    "input_token_count": 734,
    "tokens_per_second": 37.62599709057091,
    "output_token_count": 625,
    "predict_time": 16.899976448,
    "total_time": 16.910512,
    "time_to_first_token": 0.299656967
  },
  "output": [
    "<html>\n<hea",
    "d>\n  ",
    "<title>Welcome",
    " to My",
    " CYBER",
    "-",
    "SPACE",
    "</title>\n  ",
    "<style>\n    ",
    "body {\n      backgroun",
    "d:",
    " url('space",
    "-bg",
    ".jpg');",
    "\n      color",
    ": white;\n      ",
    "font-family: Arial",
    ", sans-serif;",
    "\n    ",
    "}\n    .header",
    " {\n      text-",
    "align: center;",
    "\n    }\n    .logo",
    " {\n      font",
    "-family",
    ": \"Comic",
    " Sans MS\", cursive",
    ";\n      color",
    ": #00",
    "ffff;\n    }",
    "\n    .main",
    "-content {\n      ",
    "width: 800",
    "px;\n      margin",
    ": 0 auto",
    ";\n      backgroun",
    "d: rgba(0, ",
    "0, 0,",
    " 0.7);",
    "\n      padding: ",
    "20px;\n      ",
    "border:",
    " 2px solid #",
    "4169e1;\n      ",
    "border-radius: ",
    "10px;\n    ",
    "}\n    .button",
    " {\n      backgroun",
    "d: #808080;\n      color",
    ": white;\n      ",
    "padding: 5px 15",
    "px;\n      margin",
    ": 5px;",
    "\n      border: ",
    "1px solid #666",
    ";\n      border-radius",
    ": 3",
    "px;\n    }",
    "\n    .g",
    "uestbook {\n      border",
    ": 1px soli",
    "d #4169e1;",
    "\n      padding: 10",
    "px;\n      margin",
    ": 10px 0",
    ";\n    }",
    "\n    .navigation",
    " {\n      display",
    ": flex;\n      ",
    "justify-content: center",
    ";\n      gap",
    ": 10px;",
    "\n      margin: 20",
    "px 0;",
    "\n    }\n  </style>\n</",
    "head>\n<body",
    ">\n  <div",
    " class=\"header\">",
    "\n    <div class=\"logo\">",
    "Rono!</div>\n    ",
    "<h1>Welcome to",
    "</h1>\n    ",
    "<h2>My",
    "</h2>\n    ",
    "<h2 style",
    "=\"color: red;\">CYBER-",
    "SPACE</h2>",
    "\n  </div>",
    "\n\n  <div class=\"",
    "main-content\">",
    "\n    <div class=\"navigation\">",
    "\n      <button",
    " class=\"button\">",
    "Rr",
    "pcre Sapes",
    "</button>\n      ",
    "<button class=\"button\">Art",
    " Lygp",
    "duk</button>\n      ",
    "<button class=\"button",
    "\">Sed las</button>\n    ",
    "</div>\n\n    ",
    "<div class=\"content",
    "-area\">\n      ",
    "<div class=\"guestbook",
    "\">\n        <h",
    "3>",
    "GUESTBOOK</h",
    "3>\n        ",
    "<form>\n          ",
    "<textarea rows=\"4\" cols",
    "=\"50\"></",
    "textarea>\n        ",
    "</form>\n      ",
    "</div>\n    ",
    "</div>",
    "\n\n    <div class=\"",
    "navigation\">\n      <button",
    " class=\"button\">",
    "Canay",
    "? Sy",
    "argh</button>\n      ",
    "<button class=\"button",
    "\">Apor",
    " tov daybe",
    "</button>\n    ",
    "</div>\n  ",
    "</div>\n</",
    "body>\n</html",
    ">"
  ],
  "started_at": "2025-02-12T16:20:48.270536Z",
  "status": "succeeded",
  "urls": {
    "stream": "https://stream-b.svc.ric1.c.replicate.net/v1/streams/ntlea4obxilrufuthgfy44hf673rna6u5mt7tfxngrjyjiqfgkuq",
    "get": "https://api.replicate.com/v1/predictions/vhjdxs480hrme0cmzbvbk4xc0r",
    "cancel": "https://api.replicate.com/v1/predictions/vhjdxs480hrme0cmzbvbk4xc0r/cancel"
  },
  "version": "hidden"
}

Generated in

16.9 seconds

Input tokens

730

Output tokens

620

Tokens per second

37.63 tokens / second

Time to first token

0.3 seconds

Tweak it Share

Want to make some of these yourself?

Run this model

anthropic / claude-3.5-sonnet

Prediction

Input

Output

Prediction

Input

Output

Prediction

Input

Output

Logs (8bwz4aszy5rm80cmzbttpakyfc)

Logs (p5nes79jvhrma0cmzbv8c3dx38)

Logs (vhjdxs480hrme0cmzbvbk4xc0r)