OpenAI launches Codex-Spark with 15 times faster code generation
OpenAI just launched Codex-Spark, a new AI model that generates code 15 times faster than the previous version.
Announced on February 13, it's available now for Pro users at $200 a month.
Codex-Spark is all about real-time coding with less lag, making it easier to chat and work with the AI as you go, way smoother than old-school batch programming tools.
Codex-Spark uses Cerebras WSE 3
Powered by Cerebras's Wafer Scale Engine 3 and session initialization and streaming optimizations, Codex-Spark cuts down on delays using persistent WebSocket connections.
While it's a speed champ, OpenAI says it doesn't beat its predecessor in advanced software engineering tasks and isn't top-tier for cybersecurity jobs.
Still, it's a big step toward making AI tools more flexible for both quick coding sprints and deeper problem-solving.