Skip to content
Cloudflare Docs
DirectoryAPISDKsChangelog
OpenAI logo

GPT-5

Text Generation • OpenAI • Proxied

OpenAI's model excelling at coding, writing, and reasoning.

Model Info
Context Window ↗128,000 tokens
Terms and Licenselink ↗
More informationlink ↗
PricingView pricing in the Cloudflare dashboard ↗

Usage

TypeScript
const response = await env.AI.run(
  'openai/gpt-5',
  { messages: [{ content: 'What are the three laws of thermodynamics?', role: 'user' }] },
)
console.log(response)
 
- First law: Conservation of energy. The change in a system’s internal energy equals heat added to the system minus work done by the system (ΔU = Q − W).

- Second law: Entropy of an isolated system never decreases; natural processes have a preferred direction (e.g., heat flows spontaneously from hot to cold; no heat engine is 100% efficient).

- Third law: As temperature approaches absolute zero, a system’s entropy approaches a constant minimum (often zero for a perfect crystal), and absolute zero cannot be reached in a finite number of steps.

Note: Many summaries also include the “zeroth law,” which defines temperature via thermal equilibrium.

Examples

With System Message — Using a system message to set context
TypeScript
const response = await env.AI.run(
  'openai/gpt-5',
  {
    messages: [
      { content: 'You are a helpful coding assistant specializing in Python.', role: 'system' },
      { content: 'How do I read a JSON file in Python?', role: 'user' },
    ],
  },
)
console.log(response)
 
The built-in json module does this.

Basic case (read a JSON file into a Python dict/list):
    import json

    with open('data.json', 'r', encoding='utf-8') as f:
        data = json.load(f)   # data is now a Python dict/list
    print(data)

From a JSON string (not a file):
    import json
    data = json.loads('{"name": "Ada", "age": 36}')

Common patterns:
- Access values:
    name = data.get('name')
- Iterate a list of objects:
    for item in data:
        print(item)

Error handling:
    import json, pathlib

    path = pathlib.Path('data.json')
    try:
        with path.open('r', encoding='utf-8') as f:
            data = json.load(f)
    except FileNotFoundError:
        print('File not found')
    except json.JSONDecodeError as e:
        print(f'Invalid JSON: {e}')

Newline-delimited JSON (one JSON object per line):
    import json

    with open('data.ndjson', 'r', encoding='utf-8') as f:
        for line in f:
            if line.strip():
                obj = json.loads(line)
                # process obj

Notes:
- JSON doesn’t allow comments or trailing commas.
- For very large files you may want streaming parsers like ijson.
- To write JSON back to a file: json.dump(data, f, ensure_ascii=False, indent=2).
Multi-turn Conversation — Continuing a conversation with context
TypeScript
const response = await env.AI.run(
  'openai/gpt-5',
  {
    max_completion_tokens: 8192,
    messages: [
      {
        content: 'I need help planning a road trip from San Francisco to Los Angeles.',
        role: 'user',
      },
      {
        content:
          "I'd be happy to help! The drive is about 380 miles and takes roughly 5-6 hours. Would you like suggestions for scenic routes or interesting stops along the way?",
        role: 'assistant',
      },
      { content: 'Yes, what are some good places to stop?', role: 'user' },
    ],
  },
)
console.log(response)
 
{
  "choices": [
    {
      "finish_reason": "length",
      "index": 0,
      "message": {
        "annotations": [],
        "content": "",
        "refusal": null,
        "role": "assistant"
      }
    }
  ],
  "created": 1776470774,
  "gatewayMetadata": {
    "keySource": "Unified"
  },
  "id": "chatcmpl-DVnUMlEDdAiZaWSb0PmVEIXwAQ3tp",
  "model": "gpt-5-2025-08-07",
  "object": "chat.completion",
  "service_tier": "default",
  "system_fingerprint": null,
  "usage": {
    "completion_tokens": 2048,
    "completion_tokens_details": {
      "accepted_prediction_tokens": 0,
      "audio_tokens": 0,
      "reasoning_tokens": 2048,
      "rejected_prediction_tokens": 0
    },
    "prompt_tokens": 76,
    "prompt_tokens_details": {
      "audio_tokens": 0,
      "cached_tokens": 0
    },
    "total_tokens": 2124
  }
}
Creative Writing — Longer completion for creative output
TypeScript
const response = await env.AI.run(
  'openai/gpt-5',
  {
    max_completion_tokens: 8192,
    messages: [
      {
        content: 'Write a short story opening about a detective finding an unusual clue.',
        role: 'user',
      },
    ],
  },
)
console.log(response)
 
{
  "choices": [
    {
      "finish_reason": "length",
      "index": 0,
      "message": {
        "annotations": [],
        "content": "",
        "refusal": null,
        "role": "assistant"
      }
    }
  ],
  "created": 1776470777,
  "gatewayMetadata": {
    "keySource": "Unified"
  },
  "id": "chatcmpl-DVnUPm3B4vHJc9MslDHIxYC0gLgGm",
  "model": "gpt-5-2025-08-07",
  "object": "chat.completion",
  "service_tier": "default",
  "system_fingerprint": null,
  "usage": {
    "completion_tokens": 2048,
    "completion_tokens_details": {
      "accepted_prediction_tokens": 0,
      "audio_tokens": 0,
      "reasoning_tokens": 2048,
      "rejected_prediction_tokens": 0
    },
    "prompt_tokens": 19,
    "prompt_tokens_details": {
      "audio_tokens": 0,
      "cached_tokens": 0
    },
    "total_tokens": 2067
  }
}
Streaming Response — Enable streaming for real-time output
TypeScript
const response = await env.AI.run(
  'openai/gpt-5',
  {
    messages: [{ content: 'Explain the concept of recursion with a simple example.', role: 'user' }],
    stream: true,
    stream_options: { include_usage: true },
  },
)
console.log(response)
 
Recursion is a technique where a function solves a problem by calling itself on a smaller version of the same problem, stopping when it reaches a simplest case (the base case).

Key ideas:
- Base case: a simple, direct answer that stops the recursion.
- Recursive case: reduce the problem and call the function again, moving toward the base case.

Simple example: factorial
Factorial of n (written n!) is n × (n−1) × (n−2) × … × 1, with 0! = 1.

Pseudocode:
function factorial(n):
  if n <= 1:
    return 1        // base case
  else:
    return n * factorial(n - 1)  // recursive case

How factorial(4) runs:
- factorial(4) = 4 * factorial(3)
- factorial(3) = 3 * factorial(2)
- factorial(2) = 2 * factorial(1)
- factorial(1) = 1  (base case)
Then it unwinds:
- factorial(2) = 2 * 1 = 2
- factorial(3) = 3 * 2 = 6
- factorial(4) = 4 * 6 = 24

Every recursive solution should:
- Have a clear base case.
- Make progress toward that base case each call.

Analogy: Think of nested Russian dolls. To find the smallest doll, you open a doll to get a smaller one (recursive step) until you reach the tiniest doll that can’t be opened (base case).

Parameters

frequency_penalty
numbermaximum: 2minimum: -2
max_completion_tokens
numberexclusiveMinimum: 0
max_tokens
numberexclusiveMinimum: 0
presence_penalty
numbermaximum: 2minimum: -2
response_format
stream
boolean
temperature
numbermaximum: 2minimum: 0
tool_choice
top_p
numbermaximum: 1minimum: 0

API Schemas (Raw)

Input
Output