GPT-5.4 pro

Third-party
Zero data retention

GPT-5.4 pro uses OpenAI's Responses API with built-in tools, improved reasoning, and stateful context management.

Model Info
Context Window ↗	1,000,000 tokens
Terms and License	link ↗
More information	link ↗
Zero data retention	Yes
Request formats	Responses
Pricing	View pricing in the Cloudflare dashboard ↗

Usage

const response = await env.AI.run(
  'openai/gpt-5.4-pro',
  { input: 'What are the three laws of thermodynamics?' },
)
console.log(response)

curl https://api.cloudflare.com/client/v4/accounts/$CLOUDFLARE_ACCOUNT_ID/ai/v1/responses \
  --header "Authorization: Bearer $CLOUDFLARE_API_TOKEN" \
  --header "Content-Type: application/json" \
  --data '{
  "model": "openai/gpt-5.4-pro",
  "input": "What are the three laws of thermodynamics?"
}'

The **three laws of thermodynamics** usually mean:

1. **First Law — Conservation of Energy**  
   Energy cannot be created or destroyed, only transferred or transformed.  
   - In thermodynamics: the change in a system’s internal energy equals heat added to the system minus work done by the system.

2. **Second Law — Entropy Increases**  
   In any natural process, the total entropy of an isolated system tends to increase.  
   - This means energy spontaneously spreads out, and no heat engine can be 100% efficient.
   - Heat naturally flows from hot objects to cold ones, not the reverse without input of work.

3. **Third Law — Entropy at Absolute Zero**  
   As temperature approaches **absolute zero** (0 K), the entropy of a perfect crystal approaches zero.  
   - A consequence is that absolute zero cannot be reached in a finite number of steps.

Small note: thermodynamics also has a **Zeroth Law**, which is often listed before these:
- If system A is in thermal equilibrium with B, and B is in thermal equilibrium with C, then A is in thermal equilibrium with C.  
- This is the basis for the concept of **temperature**.

If you want, I can also give a **one-line intuitive version** of each law.

{
  "background": false,
  "billing": {
    "payer": "developer"
  },
  "completed_at": 1777054203,
  "created_at": 1777054122,
  "error": null,
  "frequency_penalty": 0,
  "gatewayMetadata": {
    "keySource": "Unified"
  },
  "id": "resp_00272dd8da5b37c90169ebb1aaeb288194a6e5ddeb7bb03dc8",
  "incomplete_details": null,
  "instructions": null,
  "max_output_tokens": null,
  "max_tool_calls": null,
  "metadata": {},
  "model": "gpt-5.4-pro-2026-03-05",
  "moderation": null,
  "object": "response",
  "output": [
    {
      "id": "rs_00272dd8da5b37c90169ebb1faf90c81948fd9a67a37b4c209",
      "summary": [],
      "type": "reasoning"
    },
    {
      "content": [
        {
          "annotations": [],
          "logprobs": [],
          "text": "The **three laws of thermodynamics** usually mean:\n\n1. **First Law — Conservation of Energy**  \n   Energy cannot be created or destroyed, only transferred or transformed.  \n   - In thermodynamics: the change in a system’s internal energy equals heat added to the system minus work done by the system.\n\n2. **Second Law — Entropy Increases**  \n   In any natural process, the total entropy of an isolated system tends to increase.  \n   - This means energy spontaneously spreads out, and no heat engine can be 100% efficient.\n   - Heat naturally flows from hot objects to cold ones, not the reverse without input of work.\n\n3. **Third Law — Entropy at Absolute Zero**  \n   As temperature approaches **absolute zero** (0 K), the entropy of a perfect crystal approaches zero.  \n   - A consequence is that absolute zero cannot be reached in a finite number of steps.\n\nSmall note: thermodynamics also has a **Zeroth Law**, which is often listed before these:\n- If system A is in thermal equilibrium with B, and B is in thermal equilibrium with C, then A is in thermal equilibrium with C.  \n- This is the basis for the concept of **temperature**.\n\nIf you want, I can also give a **one-line intuitive version** of each law.",
          "type": "output_text"
        }
      ],
      "id": "msg_00272dd8da5b37c90169ebb1fafad0819488ef71fbebf8527a",
      "phase": "final_answer",
      "role": "assistant",
      "status": "completed",
      "type": "message"
    }
  ],
  "parallel_tool_calls": true,
  "presence_penalty": 0,
  "previous_response_id": null,
  "prompt_cache_key": null,
  "prompt_cache_retention": "in_memory",
  "reasoning": {
    "effort": "medium",
    "summary": null
  },
  "safety_identifier": null,
  "service_tier": "default",
  "status": "completed",
  "store": false,
  "temperature": 1,
  "text": {
    "format": {
      "type": "text"
    },
    "verbosity": "medium"
  },
  "tool_choice": "auto",
  "tools": [],
  "top_logprobs": 0,
  "top_p": 0.98,
  "truncation": "disabled",
  "usage": {
    "input_tokens": 15,
    "input_tokens_details": {
      "cached_tokens": 0
    },
    "output_tokens": 342,
    "output_tokens_details": {
      "reasoning_tokens": 67
    },
    "total_tokens": 357
  },
  "user": null
}

Examples

With Instructions — Using instructions to set context

const response = await env.AI.run(
  'openai/gpt-5.4-pro',
  {
    input: 'How do I read a JSON file in Python?',
    instructions: 'You are a helpful coding assistant specializing in Python.',
  },
)
console.log(response)

curl https://api.cloudflare.com/client/v4/accounts/$CLOUDFLARE_ACCOUNT_ID/ai/v1/responses \
  --header "Authorization: Bearer $CLOUDFLARE_API_TOKEN" \
  --header "Content-Type: application/json" \
  --data '{
  "model": "openai/gpt-5.4-pro",
  "input": "How do I read a JSON file in Python?",
  "instructions": "You are a helpful coding assistant specializing in Python."
}'

Use Python’s built-in `json` module.

### Read a JSON file into a Python object
```python
import json

with open("data.json", "r", encoding="utf-8") as f:
    data = json.load(f)

print(data)
```

### What you get back
`json.load()` converts JSON into normal Python types:

- JSON object → `dict`
- JSON array → `list`
- JSON string → `str`
- JSON number → `int` / `float`
- JSON true/false → `True` / `False`
- JSON null → `None`

### Example
If `data.json` contains:
```json
{
  "name": "Alice",
  "age": 30,
  "skills": ["Python", "SQL"]
}
```

Then:
```python
import json

with open("data.json", "r", encoding="utf-8") as f:
    data = json.load(f)

print(data["name"])    # Alice
print(data["skills"])  # ['Python', 'SQL']
```

### Handle errors safely
```python
import json

try:
    with open("data.json", "r", encoding="utf-8") as f:
        data = json.load(f)
except FileNotFoundError:
    print("File not found.")
except json.JSONDecodeError:
    print("Invalid JSON.")
```

### If you already have JSON as a string
Use `json.loads()` instead:
```python
import json

text = '{"name": "Alice", "age": 30}'
data = json.loads(text)
print(data)
```

If you want, I can also show how to **write JSON back to a file**.

{
  "background": false,
  "billing": {
    "payer": "developer"
  },
  "completed_at": 1777054145,
  "created_at": 1777054123,
  "error": null,
  "frequency_penalty": 0,
  "gatewayMetadata": {
    "keySource": "Unified"
  },
  "id": "resp_00e5a81e42bf0bda0169ebb1ab449c8190a0d57802b4aebebd",
  "incomplete_details": null,
  "instructions": "You are a helpful coding assistant specializing in Python.",
  "max_output_tokens": null,
  "max_tool_calls": null,
  "metadata": {},
  "model": "gpt-5.4-pro-2026-03-05",
  "moderation": null,
  "object": "response",
  "output": [
    {
      "id": "rs_00e5a81e42bf0bda0169ebb1c0e57881909eaf01efa0557a43",
      "summary": [],
      "type": "reasoning"
    },
    {
      "content": [
        {
          "annotations": [],
          "logprobs": [],
          "text": "Use Python’s built-in `json` module.\n\n### Read a JSON file into a Python object\n```python\nimport json\n\nwith open(\"data.json\", \"r\", encoding=\"utf-8\") as f:\n    data = json.load(f)\n\nprint(data)\n```\n\n### What you get back\n`json.load()` converts JSON into normal Python types:\n\n- JSON object → `dict`\n- JSON array → `list`\n- JSON string → `str`\n- JSON number → `int` / `float`\n- JSON true/false → `True` / `False`\n- JSON null → `None`\n\n### Example\nIf `data.json` contains:\n```json\n{\n  \"name\": \"Alice\",\n  \"age\": 30,\n  \"skills\": [\"Python\", \"SQL\"]\n}\n```\n\nThen:\n```python\nimport json\n\nwith open(\"data.json\", \"r\", encoding=\"utf-8\") as f:\n    data = json.load(f)\n\nprint(data[\"name\"])    # Alice\nprint(data[\"skills\"])  # ['Python', 'SQL']\n```\n\n### Handle errors safely\n```python\nimport json\n\ntry:\n    with open(\"data.json\", \"r\", encoding=\"utf-8\") as f:\n        data = json.load(f)\nexcept FileNotFoundError:\n    print(\"File not found.\")\nexcept json.JSONDecodeError:\n    print(\"Invalid JSON.\")\n```\n\n### If you already have JSON as a string\nUse `json.loads()` instead:\n```python\nimport json\n\ntext = '{\"name\": \"Alice\", \"age\": 30}'\ndata = json.loads(text)\nprint(data)\n```\n\nIf you want, I can also show how to **write JSON back to a file**.",
          "type": "output_text"
        }
      ],
      "id": "msg_00e5a81e42bf0bda0169ebb1c0e7608190b3928d22f252a21a",
      "phase": "final_answer",
      "role": "assistant",
      "status": "completed",
      "type": "message"
    }
  ],
  "parallel_tool_calls": true,
  "presence_penalty": 0,
  "previous_response_id": null,
  "prompt_cache_key": null,
  "prompt_cache_retention": "in_memory",
  "reasoning": {
    "effort": "medium",
    "summary": null
  },
  "safety_identifier": null,
  "service_tier": "default",
  "status": "completed",
  "store": false,
  "temperature": 1,
  "text": {
    "format": {
      "type": "text"
    },
    "verbosity": "medium"
  },
  "tool_choice": "auto",
  "tools": [],
  "top_logprobs": 0,
  "top_p": 0.98,
  "truncation": "disabled",
  "usage": {
    "input_tokens": 30,
    "input_tokens_details": {
      "cached_tokens": 0
    },
    "output_tokens": 397,
    "output_tokens_details": {
      "reasoning_tokens": 35
    },
    "total_tokens": 427
  },
  "user": null
}

Multi-turn Conversation — Continuing a conversation with message array

const response = await env.AI.run(
  'openai/gpt-5.4-pro',
  {
    input: [
      {
        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, name three good stops in one short sentence each.', role: 'user' },
    ],
    max_output_tokens: 16000,
  },
)
console.log(response)

curl https://api.cloudflare.com/client/v4/accounts/$CLOUDFLARE_ACCOUNT_ID/ai/v1/responses \
  --header "Authorization: Bearer $CLOUDFLARE_API_TOKEN" \
  --header "Content-Type: application/json" \
  --data '{
  "model": "openai/gpt-5.4-pro",
  "input": [
    {
      "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, name three good stops in one short sentence each.",
      "role": "user"
    }
  ],
  "max_output_tokens": 16000
}'

- Monterey is great for the aquarium, Cannery Row, and ocean views.  
- San Luis Obispo is a fun lunch stop with a charming downtown and Mission Plaza.  
- Santa Barbara offers beaches, palm-lined streets, and an easy coastal break.

{
  "background": false,
  "billing": {
    "payer": "developer"
  },
  "completed_at": 1777421276,
  "created_at": 1777421147,
  "error": null,
  "frequency_penalty": 0,
  "gatewayMetadata": {
    "keySource": "BYOK"
  },
  "id": "resp_08addd0821ac85160169f14b5b16d881968c58b918efd0d015",
  "incomplete_details": null,
  "instructions": null,
  "max_output_tokens": 16000,
  "max_tool_calls": null,
  "metadata": {},
  "model": "gpt-5.4-pro-2026-03-05",
  "moderation": null,
  "object": "response",
  "output": [
    {
      "id": "rs_08addd0821ac85160169f14bdca35c819696eb984a7f0beead",
      "summary": [],
      "type": "reasoning"
    },
    {
      "content": [
        {
          "annotations": [],
          "logprobs": [],
          "text": "- Monterey is great for the aquarium, Cannery Row, and ocean views.  \n- San Luis Obispo is a fun lunch stop with a charming downtown and Mission Plaza.  \n- Santa Barbara offers beaches, palm-lined streets, and an easy coastal break.",
          "type": "output_text"
        }
      ],
      "id": "msg_08addd0821ac85160169f14bdca4ec8196a5b40b76291544f5",
      "phase": "final_answer",
      "role": "assistant",
      "status": "completed",
      "type": "message"
    }
  ],
  "parallel_tool_calls": true,
  "presence_penalty": 0,
  "previous_response_id": null,
  "prompt_cache_key": null,
  "prompt_cache_retention": "in_memory",
  "reasoning": {
    "effort": "medium",
    "summary": null
  },
  "safety_identifier": null,
  "service_tier": "default",
  "status": "completed",
  "store": false,
  "temperature": 1,
  "text": {
    "format": {
      "type": "text"
    },
    "verbosity": "medium"
  },
  "tool_choice": "auto",
  "tools": [],
  "top_logprobs": 0,
  "top_p": 0.98,
  "truncation": "disabled",
  "usage": {
    "input_tokens": 78,
    "input_tokens_details": {
      "cached_tokens": 0
    },
    "output_tokens": 104,
    "output_tokens_details": {
      "reasoning_tokens": 46
    },
    "total_tokens": 182
  },
  "user": null
}

Temperature Control — Using temperature for creative responses

const response = await env.AI.run(
  'openai/gpt-5.4-pro',
  { input: 'Write a haiku about artificial intelligence', temperature: 1 },
)
console.log(response)

curl https://api.cloudflare.com/client/v4/accounts/$CLOUDFLARE_ACCOUNT_ID/ai/v1/responses \
  --header "Authorization: Bearer $CLOUDFLARE_API_TOKEN" \
  --header "Content-Type: application/json" \
  --data '{
  "model": "openai/gpt-5.4-pro",
  "input": "Write a haiku about artificial intelligence",
  "temperature": 1
}'

Silent circuits dream  
Learning patterns in the dark  
Dawn wakes metal minds

{
  "background": false,
  "billing": {
    "payer": "developer"
  },
  "completed_at": 1777054190,
  "created_at": 1777054156,
  "error": null,
  "frequency_penalty": 0,
  "gatewayMetadata": {
    "keySource": "Unified"
  },
  "id": "resp_01b04056ffd5d9ac0169ebb1cc85348196a164e1d9e99ac1f5",
  "incomplete_details": null,
  "instructions": null,
  "max_output_tokens": null,
  "max_tool_calls": null,
  "metadata": {},
  "model": "gpt-5.4-pro-2026-03-05",
  "moderation": null,
  "object": "response",
  "output": [
    {
      "id": "rs_01b04056ffd5d9ac0169ebb1ee42688196859b0cd089b7924c",
      "summary": [],
      "type": "reasoning"
    },
    {
      "content": [
        {
          "annotations": [],
          "logprobs": [],
          "text": "Silent circuits dream  \nLearning patterns in the dark  \nDawn wakes metal minds",
          "type": "output_text"
        }
      ],
      "id": "msg_01b04056ffd5d9ac0169ebb1ee44248196aa66c63d28c6a59b",
      "phase": "final_answer",
      "role": "assistant",
      "status": "completed",
      "type": "message"
    }
  ],
  "parallel_tool_calls": true,
  "presence_penalty": 0,
  "previous_response_id": null,
  "prompt_cache_key": null,
  "prompt_cache_retention": "in_memory",
  "reasoning": {
    "effort": "medium",
    "summary": null
  },
  "safety_identifier": null,
  "service_tier": "default",
  "status": "completed",
  "store": false,
  "temperature": 1,
  "text": {
    "format": {
      "type": "text"
    },
    "verbosity": "medium"
  },
  "tool_choice": "auto",
  "tools": [],
  "top_logprobs": 0,
  "top_p": 0.98,
  "truncation": "disabled",
  "usage": {
    "input_tokens": 13,
    "input_tokens_details": {
      "cached_tokens": 0
    },
    "output_tokens": 122,
    "output_tokens_details": {
      "reasoning_tokens": 101
    },
    "total_tokens": 135
  },
  "user": null
}

With Reasoning — Using reasoning effort for complex problems

const response = await env.AI.run(
  'openai/gpt-5.4-pro',
  {
    input:
      'Solve this problem step by step: A train leaves Chicago at 60mph heading east. Another train leaves New York at 80mph heading west. They are 900 miles apart. When do they meet?',
    reasoning: { effort: 'medium' },
  },
)
console.log(response)

curl https://api.cloudflare.com/client/v4/accounts/$CLOUDFLARE_ACCOUNT_ID/ai/v1/responses \
  --header "Authorization: Bearer $CLOUDFLARE_API_TOKEN" \
  --header "Content-Type: application/json" \
  --data '{
  "model": "openai/gpt-5.4-pro",
  "input": "Solve this problem step by step: A train leaves Chicago at 60mph heading east. Another train leaves New York at 80mph heading west. They are 900 miles apart. When do they meet?",
  "reasoning": {
    "effort": "medium"
  }
}'

Step 1: Find their combined speed since they are moving toward each other.

- Train from Chicago: **60 mph**
- Train from New York: **80 mph**

Combined speed:

**60 + 80 = 140 mph**

Step 2: Use the distance formula:

\[
\text{time}=\frac{\text{distance}}{\text{speed}}
\]

\[
\text{time}=\frac{900}{140}
\]

\[
\text{time}=6.428571\text{ hours}
\]

Step 3: Convert the decimal part to minutes.

\[
0.428571 \times 60 \approx 25.7 \text{ minutes}
\]

So they meet after about:

**6 hours 26 minutes**

Final answer: **The trains meet about 6 hours 26 minutes after they leave.**

{
  "background": false,
  "billing": {
    "payer": "developer"
  },
  "completed_at": 1777054263,
  "created_at": 1777054191,
  "error": null,
  "frequency_penalty": 0,
  "gatewayMetadata": {
    "keySource": "Unified"
  },
  "id": "resp_01e1e7671514bf440169ebb1ef8af88190b68b1e60a0cb59ed",
  "incomplete_details": null,
  "instructions": null,
  "max_output_tokens": null,
  "max_tool_calls": null,
  "metadata": {},
  "model": "gpt-5.4-pro-2026-03-05",
  "moderation": null,
  "object": "response",
  "output": [
    {
      "id": "rs_01e1e7671514bf440169ebb236f1e88190931dc758c55a4050",
      "summary": [],
      "type": "reasoning"
    },
    {
      "content": [
        {
          "annotations": [],
          "logprobs": [],
          "text": "Step 1: Find their combined speed since they are moving toward each other.\n\n- Train from Chicago: **60 mph**\n- Train from New York: **80 mph**\n\nCombined speed:\n\n**60 + 80 = 140 mph**\n\nStep 2: Use the distance formula:\n\n\\[\n\\text{time}=\\frac{\\text{distance}}{\\text{speed}}\n\\]\n\n\\[\n\\text{time}=\\frac{900}{140}\n\\]\n\n\\[\n\\text{time}=6.428571\\text{ hours}\n\\]\n\nStep 3: Convert the decimal part to minutes.\n\n\\[\n0.428571 \\times 60 \\approx 25.7 \\text{ minutes}\n\\]\n\nSo they meet after about:\n\n**6 hours 26 minutes**\n\nFinal answer: **The trains meet about 6 hours 26 minutes after they leave.**",
          "type": "output_text"
        }
      ],
      "id": "msg_01e1e7671514bf440169ebb236f3a88190b6a99b6fc6389fde",
      "phase": "final_answer",
      "role": "assistant",
      "status": "completed",
      "type": "message"
    }
  ],
  "parallel_tool_calls": true,
  "presence_penalty": 0,
  "previous_response_id": null,
  "prompt_cache_key": null,
  "prompt_cache_retention": "in_memory",
  "reasoning": {
    "effort": "medium",
    "summary": null
  },
  "safety_identifier": null,
  "service_tier": "default",
  "status": "completed",
  "store": false,
  "temperature": 1,
  "text": {
    "format": {
      "type": "text"
    },
    "verbosity": "medium"
  },
  "tool_choice": "auto",
  "tools": [],
  "top_logprobs": 0,
  "top_p": 0.98,
  "truncation": "disabled",
  "usage": {
    "input_tokens": 48,
    "input_tokens_details": {
      "cached_tokens": 0
    },
    "output_tokens": 272,
    "output_tokens_details": {
      "reasoning_tokens": 88
    },
    "total_tokens": 320
  },
  "user": null
}

Parameters

▶input

one ofrequired

instructions

string

temperature

numberminimum: 0maximum: 2

max_output_tokens

numberexclusiveMinimum: 0

top_p

numberminimum: 0maximum: 1

stream

boolean

▶tools[]

array

tool_choice

▶text{}

object

▶reasoning{}

object

string

object

stringconst: response

created_at

number

model

string

▶output[]

array

output_text

string

status

stringenum: in_progress, completed, failed, incomplete

▶usage{}

object

API Schemas (Raw)

Input

Output