Entity Types (drag to move • resize corner)

GPT-5 Knowledge Graph — Infographic Overview

Key characteristics, pricing, model card, safety, and implementation insights modeled in JavaScript Object Notation for Linked Data (JSON-LD).

272k
Input token limit
128k
Output token limit (incl. reasoning)
90%
Discount on cached input tokens

Get Started — Overview

Problem

Selecting the right Large Language Model (LLM) and cost profile while maintaining safety and latency remains challenging.

Solution

Leverage GPT-5 variants, token caching, and output-centric safety alongside knowledge graph-driven governance.

Core Concepts

GPT-5 Family

GPT-5 with Mini and Nano variants; reasoning levels from minimal to high.

Routing

Real-time router selects models by complexity, tools, and explicit intent.

Safe-completions

Output-centric safety reduces harmful detail while maximizing helpfulness.

Technology & Systems

Application Programming Interface (API)

Responses API supports reasoning summaries and minimal effort for speed.

Service

OpenAI GPT-5 API — global Language model API.

Software

ChatGPT — AI assistant (Web).

Challenges

56.8%
Prompt-injection attack success (red-team, gpt-5-thinking).
Latency
Invisible reasoning tokens may slow streaming.
Dual-use
Balancing detail and safety for sensitive topics.

Solutions — OPAL and Virtuoso

Knowledge Governance

Use OPAL and Virtuoso to integrate JSON-LD, enforce policies, and serve entity-linked insights to apps.

Explore Virtuoso

Operational Efficiency

Centralize pricing, routing, and safety controls; measure token spend and caching effectiveness.

Discover OPAL

Standards & Protocols

JSON-LD

JavaScript Object Notation for Linked Data for interoperable entities.

Internationalized Resource Identifier (IRI)

Stable web identifiers for entity linking.

Open Graph (OG) & Twitter Cards

POSH metadata powering previews and social sharing.

Outcomes

50%
Lower input cost vs GPT-4o for GPT-5.
90%
Input token savings via caching on repeats.
Fewer Hallucinations
Improved instruction-following and transparency.

Implementation Strategy

Retrieve GPT-5 reasoning summaries via API

  1. Prepare request — set model to 'gpt-5' and reasoning summary='auto'.
  2. Send curl — execute with your API key.
  3. Inspect response — read 'reasoning' summary.
  4. Optimize latency — set reasoning_effort='minimal'.

Implement token caching in a chat User Interface (UI)

  1. Identify cacheable segments — system prompt and prior turns.
  2. Enable caching — mark reusable tokens for discounts.
  3. Measure savings — track reuse and effective reduction.

Choose a reasoning level per task

  1. Classify task — trivial, moderate, or complex.
  2. Select effort — minimal for speed; high for complexity.
  3. Validate outputs — adjust when quality/latency is off.

Build a conversation pricing estimator

  1. Collect tokens — input, output, and reasoning.
  2. Apply prices — multiply by per-million rates.
  3. Account for caching — discount eligible inputs by 90%.

Frequently Asked Questions (FAQ)

Entity Type Explorer

WebPage

#page

GPT-5: Key characteristics, pricing and model card

CreativeWorkSeries

#series

Series: GPT-5

Organization

#publisher

Simon Willison's Weblog

OpenAI

Brand for GPT-5 family

Anthropic
Google AI
xAI

Person

#simon-willison

Author and publisher

BlogPosting

#article

Article with sections, images, and related links.

CreativeWork (Sections)

#sec-key-model

Key model characteristics

#sec-pricing

Pricing details and discounts

#sec-prompt-injection

Prompt injection notes

VideoObject

#vid-preview

Previewing GPT-5

ImageObject

#img-prompt-injection-chart #img-pelican-main #img-pelican-mini #img-pelican-nano

Product

#prod-gpt5

Input $1.25/m, Output $10/m

#prod-gpt5-mini

Input $0.25/m, Output $2.00/m

#prod-gpt5-nano

Input $0.05/m, Output $0.40/m

Offer

Each product includes input and output offers; see UnitPriceSpecification.

UnitPriceSpecification

Product Type Price (USD) Unit
GPT-5 Input 1.25 per million input tokens
GPT-5 Output 10.00 per million output tokens
GPT-5 Mini Input 0.25 per million input tokens
GPT-5 Mini Output 2.00 per million output tokens
GPT-5 Nano Input 0.05 per million input tokens
GPT-5 Nano Output 0.40 per million output tokens

FAQPage / Question / Answer

See the FAQ accordion above for interactive Q&A.

DefinedTermSet

#glossary

DefinedTerm

safe-completions reasoning tokens token caching

HowTo

HowToStep

Steps are detailed within each How-To above.

Service

#svc-openai-api

SoftwareApplication

#app-chatgpt

Dataset

#dataset-redteam

OfferCatalog

#pricing-catalog

ListItem

Catalog positions for each product in the pricing list.