# GPT 5.2 

The highest-capability tier in the GPT-5.2 family, providing extended compute for problems that demand the absolute best quality OpenAI can deliver.

- **Model ID:** `openai/gpt-5.2-pro`
- **Type:** chat
- **Providers:** openai
- **Context window:** 400,000
- **Maximum output tokens:** 128,000
- **Pricing:** $21.00/1M input tokens, $168.00/1M output tokens
- **Canonical page:** https://vercel.com/ai-gateway/models/gpt-5.2-pro

## Supported parameters

Detailed capability metadata has not been reported for this model.

## Example

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

const result = streamText({
  model: 'openai/gpt-5.2-pro',
  prompt: 'Why is the sky blue?'
})
```

## About

GPT 5.2  was released on December 11, 2025 on AI Gateway as the premium tier of the GPT-5.2 model family. It applies the maximum compute within the GPT-5.2 family to produce deeper reasoning, more thorough analysis, and higher-quality output than the standard GPT-5.2 tier.

The pro tier is designed for the subset of tasks where the standard model's capability ceiling is a binding constraint. Complex research synthesis, high-stakes analysis, expert-level content creation, and multi-factor strategic decisions all benefit from the additional reasoning depth.

With a context window of 400K tokens and the full API feature set, GPT 5.2  handles any task the GPT-5.2 family supports, it simply applies more computation to each request, producing better results on problems that merit the investment.

## What to consider

GPT 5.2  applies the most compute per request of any GPT-5.2 variant. Use it for tasks where the quality difference over standard GPT-5.2 is measurable and consequential.

Most teams pair GPT 5.2  with standard GPT-5.2 or mini models, routing only the hardest queries to the pro tier.

## When to use

### Best For

- **Mission-critical analysis:** High-stakes decisions that require the deepest available reasoning
- **Premium research synthesis:** Comprehensive analysis of complex topics with maximum thoroughness
- **Expert-level content creation:** Technical documentation, legal drafting, and precision-critical writing
- **Complex strategic planning:** Multi-factor decisions that benefit from exhaustive deliberation
- **Final-stage quality gates:** Last-pass review where maximum quality justifies premium cost

### Consider Alternatives When

- **Standard workloads:** GPT-5.2 for most tasks without the pro tier's premium
- **Cost-sensitive applications:** GPT-5 mini or GPT-5.1 instant for high-volume production
- **Specialized reasoning:** o3-pro for pure mathematical and scientific chain-of-thought
- **Speed-critical responses:** Any standard or instant-tier model when latency is the priority

## Best for

- **Expert-level content creation:** Technical documentation, legal drafting, and precision-critical writing
- **Maximum-compute GPT-5.2 tier:** Applies the most compute per request of any GPT-5.2 variant
- **Selective routing pattern:** Pairs with standard GPT-5.2 or mini models to handle only the hardest queries

## Consider alternatives

- **Standard workloads:** GPT-5.2 for most tasks without the pro tier's premium
- **Cost-sensitive applications:** GPT-5 mini or GPT-5.1 instant for high-volume production
- **Specialized reasoning:** o3-pro for pure mathematical and scientific chain-of-thought
- **Speed-critical responses:** GPT-5.2 mini or GPT-5.1 instant for lower-latency workloads

## Frequently asked questions

### How does GPT 5.2  differ from standard GPT-5.2?

It applies more compute per request, producing deeper reasoning and higher-quality output on hard problems. On simple tasks the difference may be minimal.

### When is the pro tier justified?

For high-stakes analysis, mission-critical decisions, expert-level content, and tasks where the quality difference over standard GPT-5.2 directly impacts outcomes.

### What context window does GPT 5.2  support?

400K tokens, matching the GPT-5.2 family.

### How does AI Gateway handle authentication for GPT 5.2 ?

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.

### Is GPT 5.2  slower than standard GPT-5.2?

It may take longer per request due to the additional compute. For latency-sensitive applications, route only specific high-value queries to the pro tier.

### 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-5.2-pro)
- [AI Gateway documentation](https://vercel.com/docs/ai-gateway)
- [Provider model documentation](https://platform.openai.com/docs/models/gpt-5.2-pro)
- [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.