Back to models
Chat Models

OpenAI

GPT-4o Mini

OpenAI small model for focused text and image-input tasks, with structured outputs and a 128,000-token context window.

MultimodalFast
Open AppBrowse All Models

At a glance

Know the model before you prompt.

OpenAI API specifications
Context window
128,000 tokens
Maximum output
16,384 tokens
Inputs → output
Text + Images → Text
Knowledge cutoff
October 1, 2023

API model ID: gpt-4o-mini

Capabilities & boundaries

What it supports. Where the limits are.

Tool support requires the appropriate API integration; a supported tool is not automatically active in every chat.

Supported API features and tools

  • Text and image input; text output
  • Streaming, function calling, and structured outputs
  • Predicted outputs and prompt caching
  • Web search, file search, image generation tools, code interpreter, and MCP through supported integrations
  • Fine-tuning capability, subject to current platform access restrictions

Before you choose

  • No native audio or video support; the base model produces text.
  • Fine-tuning capability is subject to platform restrictions: organizations that never fine-tuned cannot start new jobs, and inactive organizations lost access July 2, 2026. Remaining active customers lose new-job creation January 6, 2027; existing inference continues until the base model is deprecated.
  • Its 16,384-token output limit is separate from its 128,000-token context window.
  • Image generation is a separate integrated tool, not native output or proof of an enabled app feature.

Non-reasoning model

Use direct instructions and concrete examples rather than a reasoning-effort selector. GPT-4o Mini is a focused non-reasoning model; validate difficult judgments separately and route uncertain cases for review instead of forcing a confident classification.

  • Chat Completions, Responses, and Batch are supported. Tools must be supplied by the integration.
  • The documented snapshot is gpt-4o-mini-2024-07-18. API usage-tier limits and tool charges are separate considerations.
  • The Assistants API retired August 26, 2026; its historical reference entry should not guide new integrations.
  • Retirement notices for audio or transcription relatives do not announce retirement of this base model. Check exact model IDs rather than grouping similarly named models together.

Put it to work

Start with a more useful prompt.

Original examples from EZ Ai Assist. Adapt these to your task and the features available in your workspace.

Workflow 01

Triage short support messages

Use explicit categories and a fallback for uncertain records.

Classify each supplied support message as billing, access, product_question, or needs_review using the definitions below. Return the message ID, selected category, a short supporting quote, and any missing information. Choose needs_review when multiple categories fit or evidence is insufficient. Do not assign urgency or sentiment unless the sender states it explicitly.

Workflow 02

Read a photographed inventory label

Extract visible content without guessing obscured characters.

Read the inventory labels in these supplied images. Return one record per visible label with product_name, item_code, quantity, and uncertainties. Transcribe only what you can see, use null for unreadable values, and mark ambiguous characters rather than guessing. Keep labels separate and do not infer quantity from packaging size or nearby objects.

Workflow 03

Draft a concise reply from approved facts

Keep a routine response within a controlled factual boundary.

Write a reply of at most 120 words to this customer question using only the approved facts below. Answer the stated question first, preserve all dates and conditions, and ask one clarifying question if the facts are insufficient. Do not invent discounts, commitments, account status, or actions already taken. List any unresolved issue after the draft for an internal reviewer.

Developer reference

OpenAI API pricing

These are OpenAI API reference prices, not EZ Ai Assist subscription prices.

View EZ Ai Assist plans
Standard processing · USD per 1,000,000 tokens
Token typePrice
Input$0.15
Cached input$0.075
Output$0.60
  • These are Standard text-token prices for the base model. The cached-input rate is $0.075 per million eligible tokens, not a rounded $0.08.
  • Fine-tuning training and inference prices are separate and remain subject to platform access restrictions.
  • Image inputs, integrated tools, and processing tiers have additional billing rules. Calculate with the actual request usage and current provider pricing.

Common questions

A few things worth knowing.

What does Mini mean for my workflow?

It identifies a smaller model, not a guarantee that every simple-looking task is safe to automate. Use representative examples and a review category for ambiguous inputs before scaling a classification or extraction workflow.

Can GPT-4o Mini read images?

Yes, it accepts text and images and returns text. For extraction tasks, ask it to mark unreadable or ambiguous content explicitly, then compare important values against the original image.

Does this model handle live voice or transcription?

Not this base model. Audio, transcription, and realtime variants have separate IDs, capabilities, and lifecycle dates. Their documentation and retirement notices should not be applied to gpt-4o-mini.

Does structured output guarantee correct facts?

No. API structured-output support constrains the response format, not whether a statement is true. Validate source evidence, required fields, and business rules even when the response matches the requested schema.

Is fine-tuning still open to everyone?

No. Capability and access are different: self-serve training is restricted for new and inactive organizations, and remaining active customers lose new-job creation January 6, 2027. Existing fine-tuned inference follows the base model’s lifecycle.

How much text can it return?

The published maximum is 16,384 output tokens within a 128,000-token context window. Set practical response-length requirements and split large deliverables into reviewable parts.

Are the listed tool and token prices my app bill?

No. They are a direct OpenAI API reference. EZ Ai Assist subscriptions and available tools are separate. The example prompts do not activate tools that the app or integration has not enabled.

Check the source

Official documentation

Specifications and API prices checked on . Example prompts and workflow advice are editorial guidance from EZ Ai Assist.

Same provider

More from OpenAI

OpenAI

GPT-6.1 Sol

GPT-6.1 Sol is OpenAI’s reasoning model for complex coding and professional work, with image input and a 1.05M-token context window.

OpenAI

GPT-6 Sol

GPT-6 Sol is an OpenAI model for complex coding and tool-assisted workflows, with image input and configurable reasoning.

OpenAI

GPT-6 Luna

GPT-6 Luna is an efficient OpenAI model for focused, high-volume tasks, with image input, structured outputs, and configurable reasoning.

OpenAI

GPT-6 Astra

GPT-6 Astra is an OpenAI model for complex reasoning, coding, and research, with image input and a 1.05M-token context window.