OpenAI has unveiled GPT-5.3-Codex-Spark, a lightweight iteration of its advanced agentic coding tool Codex, which saw its latest model launch earlier this month. Described by the company as a "smaller version" designed for accelerated inference, Spark represents a significant step in OpenAI's physical infrastructure strategy. Powering this rapid inference is a dedicated chip supplied by its hardware partner, Cerebras, marking a deeper level of integration between the two firms.
This collaboration between Cerebras and OpenAI was first announced last month, when OpenAI disclosed a multi-year agreement valued at over $10 billion. At the time, OpenAI stated that "integrating Cerebras into our mix of compute solutions is all about making our AI respond much faster." Now, Spark is being heralded by OpenAI as the "first milestone" within this strategic partnership.
Spark, engineered for swift, real-time collaboration and "rapid iteration," leverages Cerebras’ formidable Wafer Scale Engine 3 (WSE-3). The WSE-3 stands as Cerebras’ third-generation waferscale megachip, boasting an impressive 4 trillion transistors. OpenAI characterizes this new lightweight tool as a "daily productivity driver, helping users with rapid prototyping," distinguishing its purpose from the more complex, resource-intensive tasks for which the original 5.3 model was designed. Currently, Spark is accessible in a research preview for ChatGPT Pro users within the Codex application.
Ahead of the official announcement, OpenAI CEO Sam Altman subtly teased the new model via a tweet. "We have a special thing launching to Codex users on the Pro plan later today," Altman posted, adding, "It sparks joy for me."
In its formal statement, OpenAI underscored Spark's design for achieving the lowest possible latency on Codex. The company elaborated, "Codex-Spark is the first step toward a Codex that works in two complementary modes: real-time collaboration when you want rapid iteration, and long-running tasks when you need deeper reasoning and execution." OpenAI further highlighted that Cerebras’ specialized chips are particularly adept at supporting "workflows that demand extremely low latency."
While Cerebras has operated for over a decade, the advent of the AI era has propelled it into an increasingly prominent position within the technology sector. Just last week, the company announced a successful capital raise of $1 billion, pushing its valuation to an estimated $23 billion. Cerebras has also previously signaled its intent to pursue an initial public offering.
"What excites us most about GPT-5.3-Codex-Spark is partnering with OpenAI and the developer community to discover what fast inference makes possible — new interaction patterns, new use cases, and a fundamentally different model experience," stated Sean Lie, CTO and Co-Founder of Cerebras. He concluded, "This preview is just the beginning."
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.