OpenAI запускает Codex-Spark: новый инструмент для мгновенного программирования Translation: OpenAI launches Codex-Spark: a new tool for instant programming

OpenAI has introduced a lightweight iteration of its programming tool, Codex.

The GPT-5.3-Codex-Spark is positioned as a streamlined version of the model GPT-5.3 Codex, which was unveiled in February. It is designed for faster inference. To achieve this, the startup utilized a specialized chip from its hardware partner, Cerebras.

In January, the two companies announced their collaboration.

«The integration of Cerebras into our computational solutions is aimed at significantly speeding up AI responses,» OpenAI stated at the time.

The company refers to the Spark chip as a «milestone» in their partnership.

This processor is optimized for rapid real-time performance and is based on Cerebras’ third-generation Wafer Scale Engine 3, which contains four trillion transistors.

OpenAI describes the new tool as a «daily productivity driver» that assists with quick prototyping. The original GPT-5.3 Codex is tailored for more protracted and complex assignments.

In its official statement, OpenAI emphasized that Spark is designed for minimal latency within Codex.

«Codex-Spark represents the first step toward Codex, operating in two complementary modes: real-time collaborative work, which requires fast iterations, and extended tasks necessitating deep reasoning,» the company noted.

Recall that in February 2026, OpenAI released a standalone application for the programming assistant Codex.