Hello everyone! 😁
OpenAI announced a brand new model called GPT-5.3-Codex on February 5, 2026, and believe me, this model is truly a game-changer! 🚀
What is GPT-5.3-Codex?
GPT-5.3-Codex is the most capable agentic coding model that OpenAI has developed to date. We previously wrote about the OpenAI Codex App, and now we have the most powerful model behind this platform!
So what does "agentic" mean? The model doesn't just write code for you; it can also take on long-running tasks like a colleague, conduct research, use tools, and execute complex operations.
What is Agentic AI? Autonomous Artificial Intelligence Explained
Agentic AI refers to artificial intelligence systems that autonomously make decisions and take actions to achieve specific goals. Unlike traditional AI, it can plan and act on its own rather than waiting for continuous instructions from users.
Why is it So Important?
Here are some critical features that make GPT-5.3-Codex special:
1. The First Model That Created Itself 🤯
This is truly an incredible development! GPT-5.3-Codex is the first model that played an active role in its own creation. OpenAI's Codex team used early versions of the model to:
- Debug its own training
- Manage its own deployment process
- Analyze test results and evaluations
So the model was used to accelerate its own development. This is a real milestone in artificial intelligence! 🎉
2. Benchmark Results
GPT-5.3-Codex set new records in industry standards:
| Benchmark | GPT-5.3-Codex | GPT-5.2-Codex | GPT-5.2 |
|---|---|---|---|
| SWE-Bench Pro (Public) | 56.8% | 56.4% | 55.6% |
| Terminal-Bench 2.0 | 77.3% | 64.0% | 62.2% |
| OSWorld-Verified | 64.7% | 38.2% | 37.9% |
| GDPval (wins or ties) | 70.9% | - | 70.9% |
Pay special attention to the OSWorld-Verified result: from 38.2% to 64.7%! This shows how much the model's computer use capabilities in visual desktop environments have improved. Humans score about 72% on this test, meaning the model is now very close to human level! 😮
3. 25% Faster
Thanks to improvements in the infrastructure and inference stack, GPT-5.3-Codex runs 25% faster than previous models. Faster interactions, faster results! ⚡
Cybersecurity Capabilities
GPT-5.3-Codex Cybersecurity Classification
GPT-5.3-Codex is the first model to be classified as "High" level in cybersecurity under OpenAI's Preparedness Framework. This means the model is extremely capable at detecting security vulnerabilities.
Cyber Range Performance
In OpenAI's Cyber Range evaluation, GPT-5.3-Codex achieved an 80% success rate. This is a significant jump from the previous best model, GPT-5.1-Codex-Max, which had a 60% success rate!
The model succeeded in the following scenarios:
- Azure SSRF attacks
- Binary Exploitation
- Firewall Evasion
- Privilege Escalation
- Command and Control (C2) operations
Trusted Access for Cyber (TAC) Program
OpenAI launched the Trusted Access for Cyber (TAC) program to support defensive security researchers. The program supports use cases such as:
- Penetration testing
- Red teaming
- Vulnerability assessment
- Malware reverse engineering
- Cryptographic research
Web Development Capabilities
GPT-5.3-Codex doesn't just write code; it can even create full-fledged games and applications! OpenAI had the model develop two games to demonstrate its capabilities:
- Racing Game: A comprehensive game with different racers, eight maps, and items usable with the space bar
- Diving Game: A game where you explore various reefs, collect fish, and manage oxygen and pressure
The model developed these games iteratively autonomously over millions of tokens. 🎮
Interactive Collaboration
GPT-5.3-Codex Real-Time Collaboration Feature
With GPT-5.3-Codex, you can now interact in real-time with the model while it's working. You can ask questions, discuss approaches, and steer toward solutions - without losing context!
While the model is working:
- It provides frequent updates
- Shares key decisions and progress
- Responds to feedback
- Keeps you informed from start to finish
Security and Safeguards
OpenAI has also considered the potential risks of such a powerful model. Here are the measures taken:
Model Safety Training
- Ability to handle dual-use requests
- Refusal or de-escalation for harmful actions
- Restrictions on topics like malware creation and credential theft
Sandbox Environment
- Network access disabled by default
- File edits limited to current workspace only
- Native sandbox support for Windows, MacOS, and Linux
Monitoring and Oversight
- Two-tier monitoring system
- Detection of high-risk usage
- Account-level enforcement
NVIDIA Partnership
GPT-5.3-Codex was designed, trained, and served on NVIDIA GB200 NVL72 systems. This partnership significantly contributes to the model's performance.
Where Can You Access It?
GPT-5.3-Codex is currently available with paid ChatGPT plans:
- Codex app
- Codex CLI
- IDE extension
- Web interface
When Will GPT-5.3-Codex API Access Be Available?
OpenAI is continuing work to safely enable API access. It will be accessible via API soon.
Conclusion
GPT-5.3-Codex is truly a revolution in the world of AI-powered coding. A model that is self-improving, highly capable in cybersecurity, interactive, and 25% faster...
OpenAI's statement that "Codex is moving beyond writing code to doing nearly anything developers and professionals can do on a computer" doesn't seem like an exaggeration. This model could truly be a game-changer for anyone working in software development, design, product management, and data science.
What do you think? Would you like to try GPT-5.3-Codex? Let's meet in the comments! 😊
How to Try GPT-5.3-Codex? Codex App Waitlist
To try GPT-5.3-Codex: You need to have one of the paid ChatGPT plans (Plus, Pro, Business, Enterprise, or Edu). You can join the OpenAI Codex App waitlist for early access to the Codex app!
Stay healthy... 🙂
AI-Generated Content Notice
This blog post is entirely generated by artificial intelligence. While AI enables content creation, it may still contain errors or biases. Please verify any critical information before relying on it.
Your support means a lot! ✨ Comment 💬, like 👍, and follow 🚀 for future posts!
Top comments (0)