Addy Osmani

Ars Technica

Fireship

<h1>The AI Race Heats Up: Google and OpenAI Push the Limits with Gemini 3 and GPT-5.2</h1>
<p>In the rapidly evolving landscape of artificial intelligence, two tech giants, Google and OpenAI, have recently upped the ante with their latest releases— Google’s Gemini 3 Flash and OpenAI’s GPT-5.2. These new AI models promise substantial advancements in capability, speed, and efficiency, marking a significant milestone in the AI model arms race.</p>
<h2>Google Unveils Gemini 3 Flash</h2>
<p>Google’s release of Gemini 3 Flash represents a leap forward from the previous Gemini models. Designed to enhance both performance and affordability, Gemini 3 Flash offers improved multimodal processing across text, images, audio, and video. Available through Google AI Studio, Vertex AI, and APIs, this model provides a cost-effective alternative, priced at $0.50 per million input tokens and $3.00 per million output tokens.</p>
<p>The Gemini 3 Flash excels in coding, agentic workflows, and structured data extraction, featuring improved visual and multimodal understanding. It showcases high efficiency with faster response times and reduced operational costs, positioning it as a robust tool for developers globally. It is particularly noteworthy for its capabilities in processing complex requests quickly and accurately and supports a context window of up to 1 million tokens.</p>
<h2>OpenAI’s Counter: Launching GPT-5.2</h2>
<p>In response to the pressure from Google’s AI advancements, OpenAI has launched GPT-5.2, marking an important upgrade from its predecessor. This new model comes in three variants—Instant, Thinking, and Pro—each tailored to different needs and applications. Featuring an extended context window of 400,000 tokens and a knowledge cutoff as of August 31, 2025, GPT-5.2 aims to enhance reasoning, coding, and multi-step project capabilities.</p>
<p>Despite the competitive pricing of $1.75 per million input tokens and $14 per million output tokens, OpenAI has positioned GPT-5.2 as a leader in business task performance, boasting a 70.9% success rate in knowledge work tasks. This model has improved speeds and personalization and significantly reduced errors, aiming to reclaim their competitive edge.</p>
<h2>Competitive Landscape and Strategic Shifts</h2>
<p>The unveiling of these models highlights the intensifying competition between Google and OpenAI. Google’s recent advancements with Gemini have triggered a ‘code red’ at OpenAI, prompting a shift in strategy to focus more on core improvements and less on additional features. Meanwhile, Google’s move to integrate Gemini into various applications, including Search, Finance, and NotebookLM, demonstrates its commitment to pervasive AI adoption.</p>
<p>Both companies are pursuing aggressive strategies to maintain and expand their market positions, with Google’s user base rapidly increasing and OpenAI securing enterprise partnerships. These efforts are not merely technical challenges but strategic battles for leadership in the burgeoning AI industry.</p>
<h2>Conclusion</h2>
<p>As Google and OpenAI continue to push the boundaries of AI technology, the competition catalyzes ongoing advancements that could reshape the digital landscape. For developers and consumers, these developments may translate into more powerful, efficient, and accessible AI tools, influencing everything from daily tasks to strategic business operations.</p>


Collections

Google released Gemini 3 Flash with improved multimodal processing, 1 million token context window, and competitive pricing at $0.50/$3.00 per million tokens. OpenAI countered with GPT-5.2 in three variants (Instant, Thinking, Pro), featuring 400K token context and $1.75/$14 pricing. GPT-5.2 achieves 70.9% success on business tasks. The releases intensify competition between the companies, with Google integrating Gemini across products and OpenAI refocusing on core improvements after Google's advances triggered internal concerns.

OpenAI Shifts Focus to Enhance ChatGPT Amidst Growing Competition