OpenAI Releases GPT-5.3-Codex: The Most Capable Agentic Coding Model
OpenAI's latest coding model is 25% faster than its predecessor and tops major benchmarks for complex software engineering tasks.
OpenAI has launched GPT-5.3-Codex, claiming it as the most capable agentic coding model available for complex software engineering tasks.
Performance Improvements
The new model delivers significant upgrades:
- 25% faster than GPT-5.2-Codex
- Tops benchmarks including SWE-Bench Pro and Terminal-Bench 2.0
- Enhanced agentic capabilities for autonomous coding tasks
- Better handling of complex, multi-file engineering projects
Model Retirements
Alongside this release, OpenAI has retired several older models from ChatGPT:
- GPT-4o
- GPT-4.1 and GPT-4.1 mini
- OpenAI o4-mini
- GPT-5 (Instant and Thinking)
The company reports that 99.9% of usage has shifted to GPT-5.2 and newer models.
Revenue Growth
OpenAI generated $13.1 billion in revenue in 2025 and is now targeting approximately $280 billion in revenue by 2030, with a compute infrastructure target of around $600 billion.
Ads in ChatGPT
In other news, OpenAI has begun testing ads in ChatGPT for U.S. users on Free and Go tiers. Paid plans (Plus, Pro, Business, Enterprise, Education) remain ad-free.