GPT-5.3 Codex is OpenAI's most capable agentic coding model, combining frontier software engineering performance with broad reasoning capabilities. 25% faster than GPT-5.2 with state-of-the-art SWE-Bench Pro results and strong terminal proficiency.
- Most capable agentic coding model combining frontier software engineering with broad reasoning.
- 400,000-token context window with 128,000 max output tokens and reasoning effort support (low, medium, high, xhigh).
- 25% faster than GPT-5.2 with state-of-the-art performance on SWE-Bench Pro benchmarks.
- Strong terminal proficiency and real-world computer-use skills for complex development workflows.
- Pricing is $1.75 per 1M input tokens and $14 per 1M output tokens (cached input at $0.175).
Model Information
Supported Formats
GPT-5.6 Sol is OpenAI's flagship model in the GPT-5.6 family, released July 9, 2026, for the most complex professional work spanning software engineering, long-horizon research, agentic tool use, and careful result refinement.
GPT-5.6 Terra balances capability and cost efficiency for everyday production workloads in the GPT-5.6 family, released July 9, 2026.
GPT-5.6 Luna is the fastest, most cost-optimized model in the GPT-5.6 family, released July 9, 2026, built for high-volume and latency-sensitive workloads.