GPT-5.2

OpenAI · GPT-5

Earlier GPT-5.2 generation entry retained as historical context after GPT-5.3 and GPT-5.4.

Type
language
Context
400K tokens
Max Output
128K tokens
Status
legacy
Input
$1.25/1M tok
Output
$10/1M tok
API Access
Yes
License
proprietary
reasoning coding agentic tool-use long-context instruction-following
Released February 2026 · Updated March 6, 2026

Overview

Freshness note: Model capabilities, limits, and pricing can change quickly. This profile is a point-in-time snapshot last verified on February 26, 2026.

GPT-5.2 was the next major GPT-5 generation step after the original GPT-5 release. It is no longer the newest entry in the GPT family, but it remains useful as a snapshot of how OpenAI’s lineup evolved before GPT-5.3 and GPT-5.4.

Capabilities

GPT-5.2 is strongest on multi-step reasoning, technical writing, code generation, and structured task execution where strong instruction compliance matters. It is suitable as a default model in routing setups that reserve smaller or cheaper variants for high-volume preprocessing.

Technical Details

OpenAI’s model documentation lists GPT-5.2 with a 400K context window and 128K maximum output. It supports modern tool-use and orchestration patterns in the current API stack, making it practical for long-context assistant and agent workflows.

Pricing & Access

Published API pricing (per 1M tokens):

  • Input: $1.25
  • Output: $10.00

Access is available through OpenAI API model endpoints where GPT-5 family models are exposed.

Best Use Cases

Treat GPT-5.2 as historical context or a compatibility reference if you still operate older routes. For new deployments, prefer GPT-5.3 for everyday work or GPT-5.4 for harder professional workflows.

Comparisons

  • GPT-5.3 (OpenAI): Current everyday successor in the GPT-5 family.
  • GPT-5.4 (OpenAI): Current higher-capability successor for harder workflows.
  • Claude Sonnet 4.6 (Anthropic): Contemporary alternative for broad production use.