GPT-5.3 Codex
GPTSpecialized
ThinkingTool UseVisionStructured Output
About this model
GPT-5.3-Codex is OpenAI’s most advanced agentic coding model, combining the frontier software engineering performance of GPT-5.2-Codex with the broader reasoning and professional knowledge capabilities of GPT-5.2. It achieves state-of-the-art results...
Performance Tier
Specialized
GPT-5.3 Codex is a specialized model from GPT : built for a specific domain.
Domain-specific model. Optimized for a particular task such as code generation, image creation, or web search.
Pricing
This model is included in Elosia plansHigh
High pricing. Reserve this level for tasks that demand maximum quality.
| Type | per 1M tokens |
|---|---|
| Input (prompt) | $1.75 |
| Output (completion) | $14.00 |
| Cache read | $0.175 |
Capabilities
Context Length400K
Max Output Tokens128K
TokenizerGPT
Inputtext, image, file
Outputtext
Release DateFebruary 24, 2026
Benchmarks
General Intelligence
MMLU
93%
GPQA Diamond
81%
Mathematics
MATH-500
96%
Programming
HumanEval
93%
Reasoning
IFEval
94%
Multimodal
MMMU-Pro
64%
Agentic
Terminal-Bench 2.0
77.3%
Recommended Use Cases
CodingAnalysisResearch
Strengths
- Unifies frontier coding and general reasoning in a single model
- 25% faster than GPT-5.2-Codex with comparable or better quality
- Record Terminal-Bench 2.0 score (77.3%) for real-world coding tasks
- Strong professional knowledge and broad benchmark coverage (MMLU 93%)
Limitations
- Premium pricing tier
- Primarily optimized for coding — general conversation may be less natural than GPT-5.2
Resources
This model may use your data for training