Model Selection

The Codex MCP Tool provides access to OpenAI's Codex CLI models optimized for software engineering tasks.

Available Models

Model	Context	Best For	Notes
gpt-5.2-codex	Extended	Complex multi-file edits, architecture analysis	Default, latest frontier
gpt-5.1-codex-max	Extended	Deep reasoning, complex analysis	Codex-optimized flagship
gpt-5.1-codex-mini	Standard	Quick tasks, simple queries, cost-efficient	Faster, but less capable
gpt-5.2	Extended	General reasoning, broad knowledge	Latest frontier, general purpose

How to Select a Model

Using Codex CLI Directly

# Specify model with --model flag
codex --model gpt-5.2-codex "analyze @src/**/*.ts for performance issues"
codex --model gpt-5.1-codex-max "refactor @utils.js"
codex --model gpt-5.1-codex-mini "quick review of @utils.js"

Using MCP Tool (ask-codex)

// Natural language
"use gpt-5.2-codex to analyze the entire codebase architecture"
"ask codex to solve this algorithm problem"
"quick check with gpt-5.1-codex-mini on this function"

// Direct tool invocation
{
  "name": "ask-codex",
  "arguments": {
    "prompt": "analyze @src/core for optimization opportunities",
    "model": "gpt-5.2-codex"
  }
}

Using Brainstorm Tool

{
  "name": "brainstorm",
  "arguments": {
    "prompt": "innovative features for our app",
    "model": "gpt-5.2-codex",
    "methodology": "lateral"
  }
}

Model Selection Guidelines

By Task Type

Code Review & Analysis

• Quick review: gpt-5.1-codex-mini (fast, efficient)
• Comprehensive review: gpt-5.2-codex (balanced)
• Security audit: gpt-5.2-codex (highest reliability)

Architecture & Design

• System design: gpt-5.2-codex (complex analysis)
• API design: gpt-5.1-codex-max (deep reasoning)
• Quick prototypes: gpt-5.1-codex-mini (speed)

Bug Investigation

• Complex bugs: gpt-5.2-codex (thorough analysis)
• Performance issues: gpt-5.1-codex-max (balanced)
• Simple fixes: gpt-5.1-codex-mini (quick turnaround)

Documentation

• API docs: gpt-5.2-codex (comprehensive)
• Quick comments: gpt-5.1-codex-mini (efficient)
• Architecture docs: gpt-5.2-codex (thorough)

Refactoring

• Large-scale: gpt-5.2-codex (handles complexity)
• Standard refactoring: gpt-5.1-codex-max (balanced)
• Simple cleanup: gpt-5.1-codex-mini (cost-effective)

Cost Optimization Strategies

1. Start Small, Scale Up

# Initial exploration
codex --model gpt-5.1-codex-mini "@src quick overview"

# Detailed analysis if needed
codex --model gpt-5.1-codex-max "@src comprehensive analysis"

# Deep dive for critical issues
codex --model gpt-5.2-codex "@src/critical solve complex bug"

2. Match Model to Task Complexity

// Simple tasks - use mini model
{ "prompt": "add comments", "model": "gpt-5.1-codex-mini" }

// Medium complexity - flagship model
{ "prompt": "refactor module", "model": "gpt-5.1-codex-max" }

// High complexity - latest frontier
{ "prompt": "redesign architecture", "model": "gpt-5.2-codex" }

Performance Characteristics

Response Times

• gpt-5.1-codex-mini: Fastest responses, optimized for speed
• gpt-5.1-codex-max: Balanced latency, deep reasoning
• gpt-5.2-codex: Latest frontier for complex tasks
• gpt-5.2: Variable based on task type

Reliability

• gpt-5.2-codex: Latest frontier agentic coding
• gpt-5.1-codex-max: Codex-optimized flagship
• gpt-5.1-codex-mini: Good for simple tasks
• gpt-5.2: Best for general reasoning

Setting Default Models

Configuration File

In your Codex config (~/.codex/config.toml):

[defaults]
model = "gpt-5.2-codex"

Per-Request Override

{
  "prompt": "analyze code",
  "model": "gpt-5.1-codex-mini"  // Override default
}

Model Fallback Chain

The tool uses automatic fallback when a model is unavailable:

gpt-5.2-codex → gpt-5.1-codex-max → gpt-5.2

Troubleshooting

Model Not Available

# Check available models
codex -m

# The tool will automatically fallback to available models

Slow Responses

// Switch to faster model
{
  "model": "gpt-5.1-codex-mini"
}

Best Practices

1. Start with gpt-5.1-codex-mini for initial exploration
2. Use gpt-5.1-codex-max for deep reasoning tasks
3. Reserve gpt-5.2-codex for complex, critical tasks
4. Consider gpt-5.2 for non-coding reasoning tasks
5. Monitor costs and optimize model selection

See Also

• How It Works - Understanding model integration
• File Analysis - Optimizing file references for models
• Sandbox Modes - Security with different models