OpenAI Codex Spark Breaks Speed Records With Cerebras Chip

OpenAI Codex Spark Breaks Speed Records With Cerebras Chip As AI Race Heats Up

The AI infrastructure war reached a new milestone this week as OpenAI unveiled GPT-5.3-Codex-Spark, its first production model running on Cerebras Systems' wafer-scale engine. The release marks a significant shift in how AI companies approach hardware, while competitor Anthropic responded by expanding free access to Claude's most popular features.

OpenAI's Cerebras Partnership Delivers First Product

On February 12, 2026, OpenAI announced GPT-5.3-Codex-Spark, a lightweight version of its flagship coding model designed specifically for real-time collaboration and rapid iteration. What makes this release notable is the hardware: it's the first OpenAI model running on Cerebras' Wafer Scale Engine 3 (WSE-3), which packs 4 trillion transistors onto a single chip.

The partnership between OpenAI and Cerebras was announced in January 2026, reportedly valued at over $10 billion. "Integrating Cerebras into our mix of compute solutions is all about making our AI respond much faster," OpenAI said at the time.

Key details about Codex-Spark:

•Powered by Cerebras WSE-3 chip
•Designed for ultra-low latency inference
•Available as research preview for ChatGPT Pro users
•Optimized for rapid prototyping and daily productivity tasks

"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 stated.

Anthropic Fights Back With Free Features

Just days before OpenAI's announcement, Anthropic made a strategic move to attract users frustrated with AI chatbot costs. On February 11, 2026, the company announced that free users can now access features that previously required paid subscriptions:

•File creation: Generate Excel spreadsheets, PowerPoint presentations, Word documents, and PDFs directly in conversations
•Connectors: Connect to third-party services including Slack, Asana, Zapier, Stripe, Canva, Notion, Figma, and WordPress at no cost
•Skills: Access domain-specific AI skills and create custom ones
•Extended conversations: New "compaction" feature automatically summarizes earlier context

The timing appears deliberate. Two days earlier, OpenAI began rolling out ads for free ChatGPT users. Anthropic had previously promised to keep Claude ad-free, positioning this as a competitive advantage.

What This Means for Developers

The convergence of these two developments signals several trends:

1. Hardware diversity is becoming critical: OpenAI's move away from Nvidia toward Cerebras reflects growing interest in specialized AI hardware

2. Free tiers are battlegrounds: Both companies are competing aggressively for user adoption

3. Speed matters: The emphasis on low-latency inference shows the industry moving toward real-time AI experiences

Cerebras has raised $1 billion in fresh capital at a $23 billion valuation and is reportedly preparing for an IPO. The OpenAI partnership represents a significant validation of their technology.