What are the strengths of this model?

Advanced coding capabilities Optimized by OpenAI Efficient code generation

What are the weaknesses of this model?

Not open-source Limited usage conditions Detailed specifications not public

What are the best use cases?

Efficient program development Automatic code generation Tool to assist engineers

Back to Models

OpenAIProprietary

OpenAI GPT-5-Codex-Mini

Name: OpenAI GPT-5-Codex-Mini
Author: OpenAI

OpenAI GPT-5-Codex-Mini, developed by OpenAI, is a coder model specialized in programming. It was released on November 7, 2025, and is provided under a closed license.

Parameters

Undisclosed

Context Window

License

Proprietary

Release Date

2025-11-07

API Pricing

API pricing for this model is not yet available

Strengths

・Advanced coding capabilities
・Optimized by OpenAI
・Efficient code generation

Weaknesses

・Not open-source
・Limited usage conditions
・Detailed specifications not public

Use Cases

・Efficient program development
・Automatic code generation
・Tool to assist engineers

Deep Analysis

Release Date

November 2025 (exact date varies)

Context Window

200K tokens

Input Price

$1.50 / 1M tokens

Output Price

$6.00 / 1M tokens

Architecture

o4-mini fine-tune for Codex CLI

Focus

Codex CLI workflow optimization

Throughput

~50 tok/s

Strengths

・Specialized for Codex CLI workflow — optimized for the developer command-line experience
・Lower output pricing than GPT-5-Codex ($6 vs $10 per 1M tokens)
・Fine-tuned from o4-mini for better code generation in CLI context
・Good for tight prompts and low reasoning effort workflows
・Faster iteration cycles than full-size coding models

Weaknesses

・Smaller context window (200K) compared to GPT-5-Codex (400K)
・Higher input pricing than GPT-5-Codex ($1.50 vs $1.25 per 1M tokens)
・Limited to Codex CLI workflow — not as versatile as general coding models
・Superseded by GPT-5.1 Codex Mini ($0.25/$2) which is significantly cheaper
・No published SWE-bench or other standard benchmark scores

Competitor Comparison

Model	Arena	SWE	GPQA	Price
GPT-5.1 Codex Mini	~1360	N/A	N/A	$0.25/$2.00 per 1M tokens
GPT-5-Codex	~1390	~70%	~85%	$1.25/$10 per 1M tokens
o4-mini	~1370	~55%	~78%	$1.10/$4.40 per 1M tokens
Claude Haiku 4	~1350	~45%	~72%	$0.25/$1.25 per 1M tokens

Overview

GPT-5 Codex Mini (codex-mini) is a fine-tuned o4-mini model optimized for OpenAI's Codex CLI workflow, offering faster iteration cycles for command-line code generation. At $1.50/$6.00 per 1M tokens with a 200K context window, it sits between o4-mini and GPT-5-Codex in both capability and cost. It has been largely superseded by GPT-5.1 Codex Mini which offers dramatically lower pricing.

Benchmarks & Performance

GPT-5 Codex Mini does not have published SWE-bench or Terminal-Bench scores. As an o4-mini fine-tune, its coding capability is expected to be below GPT-5-Codex (~70% SWE-bench) but above base o4-mini (~55%). The model is optimized for tight prompts and low reasoning effort, making it faster than full-size models for simple code generation tasks. Its primary advantage is in the Codex CLI workflow rather than raw benchmark performance.

Detailed Comparison

GPT-5 Codex Mini has been superseded by GPT-5.1 Codex Mini ($0.25/$2) which is 6x cheaper on input and 3x cheaper on output. Compared to GPT-5-Codex ($1.25/$10), it has lower output pricing but a smaller context window (200K vs 400K). o4-mini ($1.10/$4.40) is cheaper but not specialized for coding. For Codex CLI users, GPT-5.1 Codex Mini is the recommended replacement at much lower cost.

Community Feedback

The model was primarily used by developers working with the Codex CLI tool. Users appreciated its optimization for command-line coding workflows but found the pricing ($1.50/$6) too high compared to newer alternatives. With GPT-5.1 Codex Mini at $0.25/$2, most users have migrated. The model is now considered legacy and is primarily maintained for backward compatibility.

Use Cases

Best for: Codex CLI command-line coding, quick code completions in terminal, scaffolding and boilerplate generation, localized bugfixes with tight prompts, and code style enforcement in CLI workflows. NOT suitable for: complex multi-file refactoring, autonomous long-running tasks, large codebase analysis (200K context limit), or production coding pipelines. For new projects, use GPT-5.1 Codex Mini or GPT-5.1 Codex instead.

Latest News

Released November 2025 as part of the Codex CLI tool launch. Has been superseded by GPT-5.1 Codex Mini (November 2025) and GPT-5.1-Codex-Max (November 19, 2025). The codex-mini model ID remains available on the API for backward compatibility.

Sources

Analysis generated: 2026-05-24