Major Release OpenAI · Released
GPT-5.3 Codex
What is GPT-5.3 Codex?
GPT-5.3 Codex is a Major Release AI model from OpenAI, released on February 5, 2026. It has a 400K tokens context window, costs $1.25 per million input tokens, and scores 1406 on Chatbot Arena Elo. It is available via the company's API.
Coding-specialized variant of GPT-5.3, tuned for agentic IDE workflows.
OpenAI's specialized self-improving coding model with state-of-the-art software engineering performance
Key features
- Self-improving agentic coding
- 25% faster than GPT-5.2-Codex
- 1,000+ tokens/sec generation
- First OpenAI model flagged 'high' on cybersecurity framework
How much does GPT-5.3 Codex cost?
| Metric | Price |
|---|---|
| input per mtok | $1.25 |
| output per mtok | $10.00 |
| cached input per mtok | $0.13 |
| cache read per mtok | $0.17 |
Pricing data verified .
What context window does GPT-5.3 Codex have?
- Context window
- 400K tokens
- Max output
- 100K
- Input modalities
- text, image
- Output modalities
- text
How does GPT-5.3 Codex score on benchmarks?
| Benchmark | Score |
|---|---|
| SWE-bench Verified | 82.4% |
| HumanEval | 96.8% |
| LiveCodeBench | 84.2% |
| Aider Polyglot | 79.5% |
| GPQA Diamond | 91.5% |
| HLE | 39.9% |
| SciCode | 53.2% |
| TAU2-bench | 86.0% |
| TerminalBench-Hard | 53.0% |
| IF-Bench | 75.4% |
| LiveCodeBench Reasoning | 74.0% |
| AA Intelligence Index | 53.6 |
| Chatbot Arena Elo | 1406 |
Benchmark scores verified .
What's new
- +4pt on SWE-bench Verified vs GPT-5.2 Codex
- Native IDE tool-calling at reduced latency
- Extended max output to 100K for multi-file patches
How does GPT-5.3 Codex compare to GPT-5.2 Codex?
- SWE-bench Verified: 78.2% → 82.4% (+4.2pt)
- Input pricing: $1.50 → $1.25
- Context window: 256K tokens → 400K tokens
Compare GPT-5.3 Codex with
Launch coverage
- Openai Just Launched Gpt-5.3-Codex: A Faster Agentic Coding Model Unifying Frontier Code Performance And Professional... MarkTechPost
- Gpt-5.3-Codex System Card OpenAI
- Introducing Gpt-5.3-Codex OpenAI
- They Actually Dropped Gpt-5.3 Codex The Minute Opus 4.6 Dropped Lol Reddit
- Gpt-5.3 Codex Vs Opus 4.6: We Benchmarked Both On Our Production Rails Codebase — The... Reddit