# GPT 5.4

The standard tier of the GPT-5.4 family, bringing agentic reasoning across all domains with improved speed and token efficiency over GPT-5.2.

- **Model ID:** `openai/gpt-5.4`
- **Type:** chat
- **Providers:** openai, azure
- **Context window:** 1,050,000
- **Maximum output tokens:** 128,000
- **Pricing:** $2.50/1M input tokens, $15/1M output tokens
- **Canonical page:** https://vercel.com/ai-gateway/models/gpt-5.4

## Supported parameters

Detailed capability metadata has not been reported for this model.

## Example

```ts
import { streamText } from 'ai'

const result = streamText({
  model: 'openai/gpt-5.4',
  prompt: 'Why is the sky blue?'
})
```

## About

GPT 5.4 became available on March 5, 2026 on AI Gateway as the standard tier of the GPT-5.4 model family. It extends the agentic and reasoning leaps introduced in GPT-5.3 Codex to all domains, including knowledge work like reports, spreadsheets, presentations, and analysis.

The model handles complex multi-step workflows more reliably than previous generations, including tasks that involve tools, research, and pulling from multiple sources. It's faster and more token-efficient than GPT-5.2, delivering better results at lower cost per task.

With the context window of 1.1M tokens and the full API feature set, GPT 5.4 supports text, image, and mixed-modality inputs. If you're starting a new project or upgrading from an earlier GPT-5.x model, it's the default starting point for general-purpose work.

## What to consider

GPT 5.4 brings the agentic and reasoning leaps from GPT-5.3 Codex into all domains, not just coding.

It's faster and more token-efficient than GPT-5.2, meaning lower costs and shorter latencies on comparable tasks.

## When to use

### Best For

- **Complex multi-step workflows:** Tasks involving tools, research, and pulling from multiple sources
- **Knowledge work:** Reports, spreadsheets, presentations, and analysis across business domains
- **Advanced code generation:** Strong coding performance with GPT-5.4 generation reasoning improvements
- **Agentic applications:** Autonomous agents that coordinate tools, research, and multi-source workflows
- **General-purpose AI:** New projects that benefit from GPT-5.4 generation capability

### Consider Alternatives When

- **Cost optimization:** GPT-5.4 mini for production workloads where cost efficiency matters
- **High-volume lightweight tasks:** GPT-5.4 nano for classification, routing, and sub-agent workflows
- **Extended reasoning:** GPT-5.4 pro for maximum performance on the most complex tasks
- **Pure chain-of-thought:** o3 for mathematical and scientific reasoning tasks

## Best for

- **Complex multi-step workflows:** Tasks involving tools, research, and pulling from multiple sources
- **Knowledge work:** Reports, spreadsheets, presentations, and analysis across business domains
- **Advanced code generation:** Strong coding performance with GPT-5.4 generation reasoning improvements
- **Agentic applications:** Autonomous agents that coordinate tools, research, and multi-source workflows
- **General-purpose AI:** New projects that benefit from GPT-5.4 generation capability

## Consider alternatives

- **Cost optimization:** GPT-5.4 mini for production workloads where cost efficiency matters
- **High-volume lightweight tasks:** GPT-5.4 nano for classification, routing, and sub-agent workflows
- **Extended reasoning:** GPT-5.4 pro for maximum performance on the most complex tasks
- **Pure chain-of-thought:** o3 for mathematical and scientific reasoning tasks

## Frequently asked questions

### How does GPT 5.4 improve over GPT-5.2?

It extends the agentic and reasoning leaps from GPT-5.3 Codex to all domains. It's also faster and more token-efficient, reducing cost per task.

### What context window does GPT 5.4 support?

1.1M tokens, supporting extensive document and codebase processing.

### What types of workflows does GPT 5.4 handle well?

Complex multi-step workflows involving tools, research, and pulling from multiple sources. It also handles knowledge work like reports, presentations, and analysis.

### Does GPT 5.4 support function calling and structured outputs?

Yes. It supports the full API feature set including function calling, structured outputs, vision, and system messages.

### Should I migrate from GPT-5.2?

If your application benefits from improved agentic reasoning and you want better speed and token efficiency, yes. Test with your specific workloads to confirm the improvement.

### How does AI Gateway handle authentication for GPT 5.4?

AI Gateway accepts a single API key or OIDC token for all requests. You don't embed OpenAI credentials in your application; AI Gateway routes and authenticates on your behalf.

### What are typical latency characteristics?

This page shows live throughput and time-to-first-token metrics measured across real AI Gateway traffic.

## Links

- [Model page](https://vercel.com/ai-gateway/models/gpt-5.4)
- [AI Gateway documentation](https://vercel.com/docs/ai-gateway)
- [Provider model documentation](https://platform.openai.com/docs/models/gpt-5.4)
- [Provider pricing](https://openai.com/api/pricing/)


If you encounter incorrect information, can’t find what you need, need a capability Vercel doesn’t currently offer, or have any other feedback to report, please do it by following the instructions at https://vercel.com/feedback.md.