GPT‑5: Key Characteristics, Pricing and Model Card

Overview

GPT‑5 is OpenAI’s latest language model, featuring multiple reasoning levels, large token limits, and improved safety features like safe-completions.

Key Model Characteristics

Hybrid multi-model architecture with minimal, low, medium, and high reasoning effort modes.
Supports text and image inputs (text-only output).
Token limits: 272k input, 128k output.
Optimized for writing, coding, and health use cases.

Products & Pricing

GPT‑5

Input: $1.25/million tokens
Output: $10/million tokens

GPT‑5 Mini

Input: $0.25/million tokens
Output: $2/million tokens

GPT‑5 Nano

Input: $0.05/million tokens
Output: $0.40/million tokens

Glossary

Safe-completions: Output moderation system designed to reduce harmful content (Learn more).
Prompt Injection: Malicious alteration of model behaviour through crafted inputs (Read more).
Parallel Test Time Compute: Speed-up technique via parallel execution (More info).
Token Caching: Reuse of previously-computed tokens for cost and latency savings.

Common Questions

What is GPT‑5?

GPT-5 is OpenAI’s latest large language model with hybrid multi-model architecture, multiple reasoning levels, and optimized for key tasks.

How much does GPT‑5 cost?

Input: $1.25/million tokens, Output: $10/million tokens. Discounts up to 90% via token caching.

What are Safe-Completions?

A safety training approach that moderates responses to avoid harmful output instead of refusing outright.

Media Examples

Pelican generated by GPT‑5 — SVG Pelican from GPT‑5.