GPT-5.4
GPT-5.4 is OpenAI’s latest frontier model, unifying the Codex and GPT lines into a single system.
gpt-5.4Released 5 Mar 2026 (Mar 2026)Pricing
List prices at OpenAI, in US dollars per million tokens.
| Tier | Input / 1M | Output / 1M | |
|---|---|---|---|
| Standard | $2.50 | $15.00 | |
| Cached input 90% off input | $0.25 | 90% off input | |
| Batch 50% off | $1.25 | $7.50 | 50% off |
| Flex 50% off | $1.25 | $7.50 | 50% off |
| Priority 100% more | $5.00 | $30.00 | 100% more |
| Prompts over 272K tokens | $5.00 | $22.50 | |
| Regional endpoints 10% more | 10% more |
What it costs a month
At OpenAI list prices, 30 days.
- Chat assistant5,000 requests a day, 2K in / 0.5K out, cached$1,673
- RAG over documents2,000 requests a day, 8K in / 0.6K out, cached$1,200
- Coding agent1,000 requests a day, 40K in / 2K out, cached$2,010
- Overnight batch20,000 requests a day, 4K in / 0.4K out, batch$4,800
Where to run it
8 hosts serve GPT-5.4. The comparison is on blended price, 3 parts input to 1 part output.
| Host | In / out per 1M | Model id | |
|---|---|---|---|
OpenAILab API gpt-5.4 | $2.50 / $15.00 | gpt-5.4 | |
DatabricksCloud databricks-gpt-5-4 | $2.50 / $15.00 | databricks-gpt-5-4 | |
| $2.50 / $15.00 | global.openai.gpt-5.4 | ||
Azure OpenAICloud gpt-5.4 | $2.50 / $15.00 Retires 2 Sep 2027 | gpt-5.4 | Retires 2 Sep 2027 |
Azure AI FoundryCloud gpt-5.4 | $2.50 / $15.00 Retires 2 Sep 2027 | gpt-5.4 | Retires 2 Sep 2027 |
GMI CloudInference gpt-5.4 | $2.50 / $15.00 | gpt-5.4 | |
ReplicateInference gpt-5.4 | $2.50 / $15.00 | gpt-5.4 | |
PerplexityGateway gpt-5.4 | $2.50 / $15.00 | gpt-5.4 | |
GitHub CopilotSubscription gpt-5.4 | In the subscription | gpt-5.4 | |
ChatGPTSubscription gpt-5.4 | In the subscription | gpt-5.4 |
Show 12 more listings: regions, snapshots and variantsHide regional listings
| OpenAI | |||||||
| 2026-03-05 gpt-5.4-2026-03-05 | $2.50 / $15.00 | gpt-5.4-2026-03-05 | |||||
| Amazon Bedrock | |||||||
| Mantle openai.gpt-5.4 | $2.75 / $16.50 +10% vs headline | openai.gpt-5.4 | |||||
| us us.openai.gpt-5.4 | $2.75 / $16.50 +10% vs headline | us.openai.gpt-5.4 | |||||
| us-gov-west-1 · Mantle openai.gpt-5.4 | $3.30 / $19.80 +32% vs headline | openai.gpt-5.4 | |||||
| us-gov-east-1 · Mantle openai.gpt-5.4 | $3.30 / $19.80 +32% vs headline | openai.gpt-5.4 | |||||
| us-gov us-gov.openai.gpt-5.4 | $3.30 / $19.80 +32% vs headline | us-gov.openai.gpt-5.4 | |||||
| Azure OpenAI | |||||||
| 2026-03-05 gpt-5.4-2026-03-05 | $2.50 / $15.00 Retires 2 Sep 2027 | gpt-5.4-2026-03-05 | Retires 2 Sep 2027 | ||||
| us gpt-5.4 | $2.75 / $16.50 +10% vs headline Retires 2 Sep 2027 | gpt-5.4 | Retires 2 Sep 2027 | ||||
| eu gpt-5.4 | $2.75 / $16.50 +10% vs headline Retires 2 Sep 2027 | gpt-5.4 | Retires 2 Sep 2027 | ||||
| us · 2026-03-05 gpt-5.4-2026-03-05 | $2.75 / $16.50 +10% vs headline Retires 2 Sep 2027 | gpt-5.4-2026-03-05 | Retires 2 Sep 2027 | ||||
| eu · 2026-03-05 gpt-5.4-2026-03-05 | $2.75 / $16.50 +10% vs headline Retires 2 Sep 2027 | gpt-5.4-2026-03-05 | Retires 2 Sep 2027 | ||||
| Azure AI Foundry | |||||||
| 2026-03-05 gpt-5.4-2026-03-05 | $2.50 / $15.00 Retires 2 Sep 2027 | gpt-5.4-2026-03-05 | Retires 2 Sep 2027 | ||||
Capabilities
Core
- Streaming
- Function calling
- JSON mode
- Structured outputs
Multimodal
- Vision
- Image input
- Audio
- Image generation
- Video
Advanced
- Reasoning
- Tool calling
- MCP
- Prompt caching
- Batch API
- Embeddings
- Takes in
- ImagesTextPDFs
- Produces
- Text
- Effort settings the source lists
- nonexhigh
- Features
- Parallel tool callsPDF inputPrompt cachingTool search
Cheaper alternatives
Current models that cost less and keep most of what this one can do: tools, vision, reasoning, structured output, caching and enough context.
Same model, cheaper host
No host sells it for less than OpenAI.
Smaller OpenAI models
- 14% less
- 29% less
- 64% less
Other frontier labs
- 29% less
- 29% less
- 47% less
- 29% less
Open-weight picks
- 29% less
- 62% less
- 69% less
- 91% less
Model line
Newer and older versions, at each lab's own price.
Call it
Using OpenAI's API.
from openai import OpenAI
client = OpenAI() # reads OPENAI_API_KEY
response = client.responses.create(model="gpt-5.4", input="Hello")
print(response.output_text)curl https://api.openai.com/v1/responses \
-H "Authorization: Bearer $OPENAI_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "gpt-5.4", "input": "Hello"}'from litellm import completion
response = completion(
model="openai/gpt-5.4",
messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)Quick answers
How much does GPT-5.4 cost?
On OpenAI, GPT-5.4 costs $2.50 per million input tokens and $15.00 per million output tokens. Cached input costs $0.25 per million tokens.
Which APIs offer GPT-5.4?
GPT-5.4 runs on 8 APIs: OpenAI, Databricks, Amazon Bedrock, Azure OpenAI, Azure AI Foundry, GMI Cloud, Replicate and Perplexity.
What is GPT-5.4's context window?
1.1M tokens, with up to 128K tokens in one reply.
Is GPT-5.4 being retired?
Azure AI Foundry and Azure OpenAI retire it on 2 Sep 2027. 6 hosts keep serving it.
Is there a newer version of GPT-5.4?
Yes. GPT-5.6 is the newest model in this line.
What are cheaper alternatives to GPT-5.4?
Models that cost less and keep most of what it can do: GPT-5.3 Codex on OpenAI (14% less), Gemini 4 Carbon on Google AI Studio (29% less), Claude Sonnet 5.5 on Anthropic (29% less) and Kimi K3.1 Pro on Databricks (29% less). Savings are on a mix of 3 input tokens to 1 output token.
When was GPT-5.4 released?
OpenAI released it on 5 Mar 2026.
What is GPT-5.4's API model id?
Use gpt-5.4 with OpenAI. Other hosts use their own ids: databricks-gpt-5-4 on Databricks, global.openai.gpt-5.4 on Amazon Bedrock.
Sources
- Databricks pricing
- Amazon Bedrock pricing
- Azure OpenAI pricing
- Azure AI Foundry pricing
- GitHub Copilot pricing
- GMI Cloud pricing
- OpenAI pricing
- Perplexity pricing
- Replicate pricing
- LiteLLM model map
Raw catalog entry for gpt-5.4 on OpenAI
{
"cache_read_input_token_cost": 2.5e-7,
"cache_read_input_token_cost_above_272k_tokens": 5e-7,
"cache_read_input_token_cost_batches": 1.3e-7,
"cache_read_input_token_cost_above_272k_tokens_batches": 2.5e-7,
"cache_read_input_token_cost_flex": 1.3e-7,
"cache_read_input_token_cost_priority": 5e-7,
"input_cost_per_token": 0.0000025,
"input_cost_per_token_above_272k_tokens": 0.000005,
"input_cost_per_token_flex": 0.00000125,
"input_cost_per_token_batches": 0.00000125,
"input_cost_per_token_above_272k_tokens_batches": 0.0000025,
"input_cost_per_token_priority": 0.000005,
"litellm_provider": "openai",
"max_input_tokens": 1050000,
"max_output_tokens": 128000,
"mode": "chat",
"output_cost_per_token": 0.000015,
"output_cost_per_token_above_272k_tokens": 0.0000225,
"output_cost_per_token_flex": 0.0000075,
"output_cost_per_token_batches": 0.0000075,
"output_cost_per_token_above_272k_tokens_batches": 0.00001125,
"output_cost_per_token_priority": 0.00003,
"reasoning_contexts": [
"auto",
"current_turn",
"all_turns"
],
"reasoning_modes": [
"standard"
],
"regional_processing_uplift_multiplier_eu": 1.1,
"regional_processing_uplift_multiplier_us": 1.1,
"fedramp_uplift_multiplier": 1.1,
"supported_endpoints": [
"/v1/chat/completions",
"/v1/batch",
"/v1/responses"
],
"supported_modalities": [
"text",
"image"
],
"supported_output_modalities": [
"text"
],
"supports_function_calling": true,
"supports_native_streaming": true,
"supports_parallel_function_calling": true,
"supports_pdf_input": true,
"supports_prompt_caching": true,
"supports_reasoning": true,
"supports_response_schema": true,
"supports_system_messages": true,
"supports_tool_choice": true,
"supports_tool_search": true,
"supports_vision": true,
"supports_none_reasoning_effort": true,
"default_reasoning_effort": "none",
"supports_xhigh_reasoning_effort": true,
"supports_minimal_reasoning_effort": false,
"input_cost_per_token_above_272k_tokens_flex": 0.0000025,
"output_cost_per_token_above_272k_tokens_flex": 0.00001125,
"cache_read_input_token_cost_above_272k_tokens_flex": 2.5e-7,
"source": "https://developers.openai.com/api/docs/pricing"
}