OpenAI has released a new AI model called Codex-Spark, designed for real-time software development. This model is powered by the Cerebras Wafer-Scale Engine and runs at over 1,000 tokens per second, enabling near-instant feedback in live coding environments. Codex-Spark aims to address the limitations of agentic coding, where machines work autonomously without human supervision.
Unlike traditional AI models, Codex-Spark is designed for iterative work, allowing developers to inject direction and sensibility into their codebase. It excels at making precise edits, revising plans, and answering contextual questions about your codebase. The model has outperformed its predecessors in agentic software engineering benchmarks and can complete tasks in a fraction of the time.
OpenAI is excited to collaborate with Cerebras on this new platform capability, which brings fast inference to production. Codex-Spark is rolling out as a research preview for ChatGPT Pro users across various platforms, including the Codex app, CLI, and VS Code extension. This marks the first release in their collaboration between Cerebras and OpenAI.
The company plans to expand this ultra-fast inference capability to its largest frontier models by 2026. With Codex-Spark, developers can regain control of their software development workflow, making it an exciting milestone for AI-powered coding.
Source: https://www.cerebras.ai/blog/openai-codexspark