Back to models
Chat Models

OpenAI

GPT-5.2 Thinking

OpenAI model for document analysis, coding, and professional reasoning, available through the API as gpt-5.2.

ReasoningKnowledge work
Open AppBrowse All Models

At a glance

Know the model before you prompt.

OpenAI API specifications
Context window
400,000 tokens
Maximum output
128,000 tokens
Inputs → output
Text + Images → Text
Knowledge cutoff
August 31, 2025

API model ID: gpt-5.2

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

  • Streaming, function calling, and structured outputs
  • Text/image input and prompt caching
  • Web search and file search
  • Code interpreter and hosted shell
  • Apply patch, skills, and MCP
  • Image generation through a tool

Before you choose

  • No native audio or video support; native output is text.
  • Fine-tuning is not supported.
  • Launch-era benchmarks describe specific evaluations, not guaranteed performance in your workspace.
  • The Thinking display name does not establish the app’s selected reasoning effort.

Choose the reasoning effort

none · defaultlowmediumhighxhigh

The API reference defaults to none even though the website uses the Thinking label. Set an effort explicitly when available and test it on your own document or coding tasks. Require citations for factual conclusions and distinguish missing evidence from a negative finding.

Unsupported settings: minimal, max.

  • OpenAI maps ChatGPT’s GPT-5.2 Thinking to API model gpt-5.2, not gpt-5.2-thinking. The Instant variant is a different API model.
  • Responses, Chat Completions, and Batch are supported. Built-in tools require the appropriate Responses API integration.
  • The documented snapshot is gpt-5.2-2025-12-11. The December 2025 launch article is historical context; use the current model reference for API limits.

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

Synthesize a document trail

Reconstruct a decision from dated evidence without filling gaps.

Using these dated documents, reconstruct how this decision changed over time. Create a timeline with the decision, supporting source and location, newly introduced constraint, and any unresolved contradiction. Distinguish a proposal from an approved decision. Then write a concise current-state summary using only the supplied evidence and list the missing records that could change it.

Workflow 02

Investigate a cross-file inconsistency

Compare contracts and behavior across a small set of related files.

Compare these related source files and API contracts for inconsistent handling of the same field. Trace validation, transformation, persistence, and response formatting only where the supplied code supports it. Cite each mismatch and a concrete input that could expose it. Propose the smallest coordinated fix and tests for each boundary. Clearly label anything that needs runtime verification.

Workflow 03

Read a chart with its supporting data

Use the visual and the supplied numbers as separate sources of evidence.

Review this chart together with its source table and explanatory notes. Identify the main supported trend, any mismatches between the visual and the data, and limitations caused by scale, missing values, or sampling. Cite the relevant row, column, or chart region. Do not infer causation. Finish with three questions a reviewer should ask before using the chart in a decision.

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$1.75
Cached input$0.175
Output$14.00
  • These are Standard rates for API model gpt-5.2, not GPT-5.2 Pro or a ChatGPT subscription.
  • The reviewed pricing table lists no separate long-context rates for gpt-5.2. Cached pricing requires eligible cached tokens.
  • Other service tiers and tools have separate pricing. Confirm current rates before estimating usage; historical launch prices alone are not a current billing reference.

Common questions

A few things worth knowing.

Is GPT-5.2 Thinking a separate API identifier?

OpenAI’s naming table maps ChatGPT-5.2 Thinking to GPT-5.2 in the API, using gpt-5.2. The site keeps its existing Thinking display name and route. Do not use an invented gpt-5.2-thinking API identifier.

Does Thinking mean reasoning is always enabled?

No. The API model reference lists none as default, with low, medium, high, and xhigh also supported. A display name does not tell you which effort an application sends. Check the actual integration or available app controls.

Is this the same as GPT-5.2 Pro or Instant?

No. OpenAI distinguishes Thinking, Pro, and Instant in its launch article. This guide describes gpt-5.2. Do not transfer Pro’s pricing or the Instant variant’s lifecycle to this API model.

Are the launch article’s ChatGPT features available here?

ChatGPT product features are not a promise of EZ Ai Assist feature parity. The launch article discusses a particular product experience and evaluations. Check the app for supported tools, uploads, output formats, and current model access.

Can it analyze charts and screenshots?

Image input is supported. Supply the original data where possible, ask for visible evidence, and verify numerical claims. A screenshot cannot confirm hidden interactions, omitted data, or causal explanations.

Are these API rates part of my subscription?

No. The rates describe direct OpenAI API usage for gpt-5.2. EZ Ai Assist subscriptions are separate, and the app determines which features and usage allowances your plan includes.

How should I use the historical benchmarks?

Treat them as reported results under the launch evaluation’s conditions, not a forecast for your own work. Build a small evaluation using your documents, expected answers, and error tolerance before choosing a model for repeated use.

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.