GPT-5.2 Codex

GPT-5.2 Codex is an specialized iteration of the GPT-5.2 architecture, purpose-built to serve as an advanced agentic tool for professional software engineering. It is designed to handle complex, real-world development workflows, offering significant improvements in managing long-horizon tasks. By integrating features like context compaction, the model excels at maintaining coherence during extensive code migrations and large-scale refactors, making it a robust assistant for developers navigating intricate codebases and Windows-based environments.

Developed through targeted optimization for agentic coding, this model builds upon the broader GPT-5.2 lineage to deliver enhanced reasoning and context understanding. Its practical strengths are particularly evident in specialized domains such as cybersecurity, where it supports security researchers in identifying vulnerabilities at scale. As an agent-focused model, it is engineered to support both developers and security defenders in tackling multi-step technical challenges, representing a forward-looking step in how AI can be applied to maintain and secure modern software infrastructure.

Azuregpt-5.2-codexgpt-codex

Open provider page Provider docs

Quick Info

Provider: Azure
Model key: gpt-5.2-codex
Release date: Jan 14, 2026
Last updated: Jan 14, 2026
Knowledge cutoff: 2025-08-31
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $1.75
Output token cost: $14.00

Limits

Input tokens: 272,000 tokens
Output tokens: 128,000 tokens
Context window: 400,000 tokens

Transparent token rates

Compare GPT-5.2 Codex pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about GPT-5.2 Codex

Abacus

CoverageBenchmark

This Magica model directory page lists GPT-5.2-Codex as developed by OpenAI with a context window of 400K tokens, an output limit of 128K tokens, and a release date of January 14, 2026. The page reports pricing at $1.75 per million input tokens, $14.00 per million output tokens, $0.01 per web search, and $0.175 per mil The page also reports a Best Design Arena Score of 1,176 Elo, placing GPT-5.2-Codex 19th in the Android Native category with a 47.5% win rate, sourced from arena comparisons. According to the supplied excerpt, the page includes comparison widgets pairing GPT-5.2-Codex against models such as GPT-5.6 Luna, GLM 5.3 Flash,

OrcaRouter

Coverage

A third-party model registry entry for GPT-5.2 Codex lists it as a closed-source, proprietary dense model with a 400K context length, text modality, and a stated July 2025 knowledge cutoff. It frames the model as specialized for autonomous multi-file refactoring, visual UI debugging, and secure system design, with pers The same page reports LiveBench category scores including Coding 0.84, Data Analysis 0.78, Global 0.74, Agentic Coding 0.49, Mathematics 0.89, Average 0.74, Reasoning 0.78, and a WebDev Arena score of 1338, giving an overall rank of 84 and a Coding rank of 37. These benchmark figures have unverified provenance and meth