OpenAI launches GPT-5.3-Codex-Spark on Cerebras chips — marks AI giants first production deployment away from Nvidia

Source: Tom’s Hardware

Image credit: OpenAI

Release Overview

OpenAI on Thursday released GPT‑5.3‑Codex‑Spark, its first AI model served on chips from Cerebras Systems. This marks the ChatGPT maker’s first production deployment on silicon outside its long‑standing core stack with Nvidia. The new model is a streamlined, lower‑power variant of Codex designed for fast, interruptible coding tasks, and it is initially rolling out as a research preview to ChatGPT Pro subscribers.

Performance Claims

• Optimized for interactive development workflows such as editing specific sections of code and running targeted tests.
• Engineered for high throughput on ultra‑low latency hardware, with OpenAI claiming the model can exceed 1,000 tokens per second under the right configuration.
• Defaults to minimal edits and will not automatically execute tests unless explicitly instructed.

Partnership and Infrastructure

OpenAI announced last month that it had signed a deal to deploy Cerebras hardware for low‑latency inference. The company plans to bring 750 MW of Cerebras‑backed compute online in phases through 2028. While this capacity will not replace Nvidia’s role in OpenAI’s training infrastructure, it provides a dedicated tier optimized for responsiveness rather than large‑scale training.

Relationship with Nvidia

Earlier this month, Sam Altman posted on X.com that OpenAI “loves working with Nvidia” and that “they make the best chips in the world,” adding, “We hope to be a gigantic customer for a very long time.” This statement followed a Reuters report suggesting OpenAI was unsatisfied with some Nvidia chips.

OpenAI continues to describe its partnership with Nvidia as foundational, noting that its most powerful models remain trained and served on Nvidia systems. At the same time, the company is expanding its ecosystem through collaborations with Cerebras and other partners.