OpenAI introduces 'Ultrafast,' a new mode that makes GPT-5.6 Sol work at 14x the speed

OpenAI, a leader in artificial intelligence, has just unveiled a significant advancement for its latest and most powerful language model, GPT-5.6 Sol. The company is introducing a new "Ultrafast" mode, a sped-up version designed specifically to meet the demanding needs of enterprise users. This new capability promises to make GPT-5.6 Sol operate at an astonishing 14 times the speed, generating up to 750 output tokens per second.

This move is a clear signal of OpenAI's intent to solidify its position in the competitive enterprise AI market. By offering unprecedented speed alongside cutting-edge intelligence, OpenAI aims to enable a new generation of real-time AI applications for businesses across various sectors.

The Quest for Real-Time AI: Why Speed Matters for Businesses

In the world of enterprise technology, speed is often as crucial as intelligence. While AI models have grown incredibly capable, the time it takes to generate responses – known as latency – has been a limiting factor for many real-time applications. Imagine a customer support chatbot that takes several seconds to formulate a reply, or an incident response system that lags in providing critical analysis. These delays, even if minor, can significantly impact efficiency, user experience, and ultimately, a company's bottom line.

For businesses, the ability to process information and generate insights instantly can unlock immense value. This is particularly true for use cases that require immediate feedback and rapid iteration. Industries like finance, cybersecurity, and customer service have a critical need for AI that can keep pace with human interaction and fast-evolving situations. OpenAI's "Ultrafast" mode directly addresses this challenge, aiming to remove the bottleneck of waiting for AI models to "think" before they type.

Introducing Ultrafast Mode for GPT-5.6 Sol

At the heart of this announcement is the "Ultrafast" mode, a new service tier for OpenAI's GPT-5.6 Sol model. GPT-5.6 Sol itself is a next-generation model known for its stronger capabilities across coding, science, and cybersecurity, alongside an advanced safety stack.

The "Ultrafast" mode boosts GPT-5.6 Sol's processing speed by up to 14 times compared to its standard operation. This translates to an impressive output of up to 750 tokens per second. To put this into perspective, benchmarks conducted by Cerebras show that GPT-5.6 Sol in Ultrafast mode runs 11 times faster than Fable 5 and 5 times faster than Opus 4.8 on Fast mode. On the GDP-Val benchmark, which measures economically valuable knowledge work tasks, Ultrafast delivered a 5.6x end-to-end speedup without any degradation in quality.

This significant leap in speed means that complex requests can be processed almost instantly, transforming how businesses can deploy and interact with AI. The mode is designed to deliver "frontier intelligence" to products and workflows where every second truly counts, without compromising on the quality of the output.

The Cerebras Partnership: Powering the Speed Revolution

Achieving such a dramatic increase in speed didn't happen in a vacuum. OpenAI's "Ultrafast" mode is powered by a strategic partnership with Cerebras, a company known for its specialized AI hardware. Specifically, Cerebras' wafer-scale processors are instrumental in enabling this breakthrough.

The technical innovation lies in how these processors handle model weights. By keeping the model weights in on-chip SRAM (Static Random-Access Memory), Cerebras' architecture significantly reduces data transport bottlenecks. This direct access to memory eliminates the delays typically associated with moving data between different chips, which is a common limitation in traditional AI inference systems. This structural judgment emphasizes how wafer-scale architecture replaces inter-chip data transport with on-chip memory, directly rewriting the upper limit of inference latency.

This collaboration highlights a growing trend in the AI industry: the importance of specialized hardware to push the boundaries of AI performance. As AI models become more complex and the demand for real-time applications grows, the synergy between advanced software models and purpose-built hardware becomes increasingly vital.

Targeting the Enterprise: Use Cases and Availability

The primary motivation behind the "Ultrafast" mode is to court enterprise users. OpenAI envisions this rapid processing capability transforming a wide array of business operations. The company has identified several key areas where this enhanced speed will be particularly impactful:

• Real-time Voice Applications: Enabling more natural and responsive AI assistants for customer service, virtual meetings, and interactive voice response (IVR) systems.
• Customer Support: Providing immediate, accurate responses to customer queries, improving satisfaction and reducing resolution times.
• Coding and Design Agents: Accelerating development cycles by providing instant code suggestions, bug fixes, and design iterations.
• Financial Research and Analysis: Enabling rapid processing of market data, financial models, and reports, crucial for high-frequency trading and timely investment decisions.
• Incident Response and Cybersecurity: Facilitating quicker analysis of security threats, log data, and automated response mechanisms, where every second can prevent significant damage.
• E-commerce: Enhancing personalized shopping experiences, dynamic pricing, and real-time inventory management.

The "Ultrafast" mode is currently rolling out as a preview to a select group of customers through the OpenAI API. OpenAI plans to expand access over time as capacity increases, indicating a strategic, phased deployment to ensure stability and gather feedback from early adopters. This initial limited release underscores the advanced nature of the technology and its enterprise-focused application.

Broader Implications for the AI Landscape

OpenAI's introduction of "Ultrafast" mode for GPT-5.6 Sol carries significant implications for the broader AI industry:

Intensifying Competition in Enterprise AI

This move intensifies the competition among major AI developers vying for the lucrative enterprise market. Companies like Google, with its Gemini 3.7 Flash model also tuned for speed in coding and business workflows, are constantly pushing the boundaries of what's possible. OpenAI's focus on raw speed with "Ultrafast" mode sets a new benchmark, forcing competitors to innovate further in low-latency performance.

New Frontiers for AI Applications

By making AI models significantly faster, OpenAI is effectively opening up new application frontiers. Many tasks previously deemed too time-sensitive for AI are now becoming viable. This could lead to a surge in innovative AI-powered products and services that rely on real-time interaction and immediate decision-making.

The Convergence of Hardware and Software

The partnership with Cerebras highlights the increasing importance of specialized hardware in achieving peak AI performance. As models grow in complexity, the efficiency of the underlying computing infrastructure becomes critical. This could spur more collaborations between AI software companies and hardware manufacturers, leading to more integrated and optimized AI solutions.

Redefining "Intelligence" with Speed

The concept that "once speed is fast enough, it becomes a new form of intelligence" is gaining traction in the AI community. The ability to respond instantaneously can fundamentally change how humans interact with AI, making the experience feel more natural, intuitive, and truly assistive. This shift could redefine user expectations and the very definition of what constitutes a "smart" AI system.

What's Next for OpenAI and Enterprise AI?

The launch of "Ultrafast" mode is a strong indicator of OpenAI's strategic direction. The company is not only focused on developing more intelligent models but also on making them practical and performant for high-stakes business environments. As the preview expands and more enterprises gain access, the real-world impact of this speed enhancement will become clearer.

This development also suggests that future OpenAI models might continue to prioritize both intelligence and efficiency, offering different tiers or modes tailored to specific performance requirements. The ongoing evolution of models like GPT-5.6 Sol, alongside its siblings Terra and Luna, demonstrates OpenAI's commitment to providing a versatile suite of AI capabilities.

For businesses considering integrating advanced AI into their operations, "Ultrafast" mode presents a compelling proposition. It promises to bridge the gap between AI's immense potential and the practical demands of real-time enterprise workflows, paving the way for more responsive, efficient, and impactful AI deployments.

Frequently Asked Questions

What is OpenAI's "Ultrafast" mode?

"Ultrafast" mode is a new service tier for OpenAI's GPT-5.6 Sol model that significantly boosts its processing speed, making it up to 14 times faster than normal operation. It can generate up to 750 output tokens per second.

Which OpenAI model uses "Ultrafast" mode?

The "Ultrafast" mode is specifically introduced for OpenAI's flagship model, GPT-5.6 Sol.

How does "Ultrafast" mode achieve its speed?

The enhanced speed is a result of a partnership with Cerebras, utilizing their wafer-scale processors. These processors keep model weights in on-chip SRAM, which helps reduce data transport bottlenecks and allows for ultra-low latency inference.

Who is the "Ultrafast" mode for?

"Ultrafast" mode is primarily aimed at enterprise users who need real-time AI capabilities for critical applications. This includes fields like customer support, financial analysis, coding, incident response, and cybersecurity, where low latency is crucial.