OpenAI unveils GPT-5.3-Codex-Spark, code 15x faster for Pro at $200/month
OpenAI just dropped GPT-5.3-Codex-Spark, a coding model that generates code 15 times faster than before. It's built for real-time edits and speedy task completion, perfect if you want to get things done quickly.
Right now, it's only for Pro users at $200 a month, with Plus users likely next, and other tiers possibly following soon.
Codex-Spark 80% less delay, weaker cybersecurity
Codex-Spark seriously cuts down lag by reducing server delays by 80% and getting your first response out twice as fast.
The trade-off? It has weaker cybersecurity capability and isn't as strong on tough coding challenges, and users may see slower access or temporary queuing during busy hours.
Still, its focus on real-time interaction hints at even smoother workflows in future updates, including a persistent WebSocket connection.