Log in
← All models

GPT-5.3-Codex

Balanced

OpenAI's coding-tuned model — the default behind Codex CLI's local runs.

OpenAI
Maker
89/100
Overall
47
Value score
$5.42
Blended /1M
400K
Context
Proprietary
Weights
Pricing: $1.75 in · $14 out / 1M tokens · synced from OpenRouter · updated 2026-06-22

Scorecard

Output Quality8.8/10 · 76% community
Agentic Ability9.3/10 · 79% community
Speed7.8/10 · 72% community
Value for $7.8/10 · 71% community
Reliability8.7/10 · 76% community

GPT-5.3-Codex is OpenAI's variant tuned specifically for software engineering: sandboxed execution, diff generation, and long agentic coding sessions. It's the model Codex CLI leans on, with a 400K context aimed at real repositories.

Strengths
  • Tuned for coding and tool use
  • Strong on long agentic sessions
  • 400K context
Trade-offs
  • Narrower than a general flagship outside coding
  • Mid-tier pricing

Agents that run GPT-5.3-Codex well

Codex CLI users and dedicated coding agents.

Best model for CodexBest model for ClineBest model for AiderBest model for OpenCodeBest model for Kilo Code

GPT-5.3-Codex FAQ

Is GPT-5.3-Codex good for AI agents?
Codex CLI users and dedicated coding agents. On our editorial scoring it rates 9.3/10 for agentic ability and 8.8/10 for output quality. Tuned for coding and tool use.
How much does GPT-5.3-Codex cost?
Via OpenRouter, GPT-5.3-Codex is priced at $1.75 in · $14 out / 1M tokens — a blended rate of about $5.42 per 1M tokens for typical input-heavy agent use.
What is GPT-5.3-Codex's context window?
GPT-5.3-Codex has a 400K-token context window. OpenAI is the maker.
Can I self-host GPT-5.3-Codex?
No — GPT-5.3-Codex is a proprietary, hosted-only model from OpenAI. You access it through an API (e.g. OpenRouter) rather than running the weights yourself.
What are the downsides of GPT-5.3-Codex?
The main trade-offs: narrower than a general flagship outside coding; mid-tier pricing. It's strongest for codex cli users and dedicated coding agents.

GPT-5.3-Codex compared

GPT-5.3-Codex vs Sonnet 4.6GPT-5.3-Codex vs GPT-5.4GPT-5.3-Codex vs Gemini 3.5 FlashGPT-5.3-Codex vs Kimi K2.7 CodeGPT-5.3-Codex vs MiniMax M3GPT-5.3-Codex vs Qwen3.7 PlusGPT-5.3-Codex vs Devstral 2GPT-5.3-Codex vs GLM 4.7

The model is half the story — the agent is the other half

The model picks the moves; the agent runs the loop, the tools, and the guardrails. Once you've chosen a model, see which agent gets the most out of it.

Compare AI agents →