FrontierThe story, in brief
Introducing GPT-5.3-Codex-Spark
15x faster. 128k context. OpenAI's new Codex-Spark is built for real-time coding—and it's already in the wild.
The KeyNews take
Why it matters
OpenAI is shipping a specialized coding model optimized for speed and context, signaling a shift toward domain-specific model releases over general-purpose scaling. This matters for founders building on OpenAI's API and investors tracking the competitive coding-AI landscape.
The key facts
7 to knowModel name: GPT-5.3-Codex-Spark
15x faster generation vs. predecessor
128k context window
Real-time coding capability (first for OpenAI)
Research preview launched
Limited to ChatGPT Pro users (initial rollout)
Published: Feb 12, 2026
Go to the source
OpenAI Blogopenai.com
Publisher excerpt: Introducing GPT-5.3-Codex-Spark—our first real-time coding model. 15x faster generation, 128k context, now in research preview for ChatGPT Pro users.