This article was generated by AI and cites original sources.
OpenAI has introduced a new iteration of its Codex tool, Codex-Spark, a lightweight version optimized for fast inference. This release marks OpenAI’s collaboration with chipmaker Cerebras, integrating a dedicated chip for enhanced performance.
The partnership with Cerebras, valued at over $10 billion, aims to accelerate AI processing speed. Codex-Spark, powered by Cerebras’ Wafer Scale Engine 3 with 4 trillion transistors, prioritizes real-time collaboration and rapid prototyping over heavier tasks.
Codex-Spark is positioned to offer low latency and support both swift collaboration and extensive reasoning within the Codex environment.
Source: TechCrunch