PDF/Image source
AI Models & Assistants / Knowledge library
GPT-5 Chat
GPT-5 model used in ChatGPT
Why learn GPT-5 Chat?
Offers the documented OpenAI model capabilities through hosted.
The problem it solves
Provides model inference for supported text / image inputs.
Where it came from
GPT-5 Chat: GPT-5 model used in ChatGPT Published aliases: gpt-5-chat-latest. Input: text, image. Output: text. Checked 2026-10-13. Model capabilities are distinct from tools available in a particular agent.
Build your understanding
Work through these exercises in order. Try the task before opening a reference.
- Explain what GPT-5 Chat does.
- Identify a use case for GPT-5 Chat.
- Describe a limitation of GPT-5 Chat.
- Name a compatible tool for GPT-5 Chat.
- Compare GPT-5 Chat against an alternative.
Documentation and sources
- Primary source / model card (developers.openai.com)
Catalog research notes
- See the AI Workbench for source dates, pricing evidence and configuration details.
Connect the ideas
Built on
Real World Example
Automated Architectural Compliance Auditing
Global engineering firms managing multi-disciplinary construction blueprints.
The business problem
Manual verification of thousands of pages of architectural blueprints against local zoning codes is prohibitively expensive and prone to human oversight errors.
How this topic solves it
Utilizing the multimodal capabilities of gpt-5-chat-latest to ingest architectural schematics as image inputs and cross-reference them against textual zoning regulations via the Chat Completions API.
- Convert high-resolution blueprint PDFs into image formats compatible with the model input modalities.
- Structure zoning code documents into a queryable format for the model context.
- Send combined image and text prompts to the v1/chat/completions endpoint using the gpt-5-chat-latest model ID.
- Parse the structured output to flag specific compliance violations for human review.
The flow of data from raw blueprints to compliance reports.
Textual constraints
Multimodal analysis engine
Flagged violations
- Blueprint Data → image_input → GPT-5 Chat
- Zoning Rules → text_context → GPT-5 Chat
- GPT-5 Chat → structured_output → Compliance Report
Implementation example
This demonstrates the use of the Chat Completions API to process multimodal inputs as required for the architectural auditing scenario.
client.chat.completions.create(model='gpt-5-chat-latest', messages=[{'role': 'user', 'content': [{'type': 'image_url', 'image_url': {'url': 'blueprint_url'}}, {'type': 'text', 'text': 'Check this against zoning code.'}]}]) # Sends image and text to the modelExpected business outcome
A significant reduction in the time required for initial compliance screening, allowing human architects to focus on complex design challenges rather than routine regulatory checks.
Limitations and checks
The model may experience hallucinations regarding precise spatial measurements in complex blueprints; verify output by checking against the model's confidence scores or implementing a secondary programmatic validation layer.