LLM ReferenceLLM Reference

GPT-5.3-Codex

gpt-5.3-codex

Proprietary

About

Most capable agentic coding model from OpenAI. Optimized for long-horizon, agentic coding tasks in the Codex CLI and API. Note: GPT-5.3-Codex-Spark is a distinct ChatGPT Pro research preview (not API-accessible).

GPT-5.3-Codex has a 400K-token context window.

GPT-5.3-Codex input tokens at $1.75/1M, output at $14/1M.

Capabilities

VisionMultimodalReasoningFunction CallingTool UseStructured OutputsCode ExecutionPrompt CachingBatch APIAudioFine-tuning

Providers(2)

Compare all →
ProviderInput (per 1M)Output (per 1M)Type
OpenRouter$1.75$14Serverless
OpenAI API$1.75$14.00Serverless

Benchmark Scores(3)

BenchmarkScoreVersionSource
SWE-bench Verified85.0SWE-bench Verifiedhttps://www.swebench.com/verified.html
SWE-bench Pro56.8—DAT-1724 research (OpenAI announcement via third party)
τ-bench77.8τ-benchhttps://taubench.com/

Rankings

Best LLMs for Code GenerationBest AI Agents & Agentic ModelsBest Multimodal / Vision LLMsBest LLMs for Reasoning & MathBest LLMs for Function Calling & Tool UseCheapest LLM APIsBest Long Context LLMsBest LLM APIsBest LLMs for EnterpriseBest LLMs for Customer Support

Compare

All comparisons →
GPT-5.3-Codex vs StepFun Step-2GPT-5.3-Codex vs Phi-4 Reasoning Vision 15B

Specifications

FamilyCodex
Released2026-02-05
Context400K
Max output128,000
ArchitectureDecoder Only
Knowledge cutoff2025-08
Specializationcode
LicenseProprietary

Created by

OpenAIOpenAI

Cutting-edge research and development.

San Francisco, California, United States
Founded 2015
Website

Providers(2)

OpenRouterOpenAI API

Links

Website