Google unveils another new AI model in three weeks, intensifying the AI market’s race for speed [Global AI Briefing]

[Financial News] Google has unveiled two new AI models: Gemini 3.8 Flash and the cybersecurity-focused Gemini 3.8 Flash Cyber. The announcement came just three weeks after the launch of Gemini 3.7 Flash, making it Google’s third Flash model released in the past six weeks.
Global AI companies, including Google, OpenAI and Anthropic, are rapidly shortening the intervals between new model launches. They are stepping up the race to develop and release AI models faster.
On the 2nd, local time, Google DeepMind announced on its official blog that it had officially launched the new language model Gemini 3.8 Flash and its cybersecurity-focused version, Gemini 3.8 Flash Cyber.

Google described the model, internally code-named “Schemakey,” as “the most intelligent Flash model.” It emphasized that the model offers significantly improved capabilities in multi-step problem-solving and coding.
In other words, the model is focused on software engineering, autonomous agent tasks and multistep reasoning.
It supports multimodal inputs, including text, images, video, audio and PDFs. The cost is $0.75 per 1 million input tokens and $3.75 per 1 million output tokens, making it a midrange, lower-cost model priced the same as Gemini 3.7, which was released three weeks ago.
The launch of Gemini 3.8 has drawn industry attention because it highlights the intensifying competition among global AI companies to release new models at ever-faster speeds.
Google CEO Sundar Pichai previously declared the company’s intention to accelerate its AI model launches. During an earnings call last July, he said, “We plan to increase the pace of AI model releases and launch models almost every month.” In fact, Google’s Gemini Flash line released version 3.6 on July 21, followed by version 3.7 on August 13 and version 3.8 in September.
OpenAI is also joining the race to release new AI models faster. After launching GPT-5.3-Codex in February, it introduced five GPT-5.x variants in succession over the following three months through May. In particular, it took less than two months to move from GPT-5.4 to GPT-5.5.
Anthropic has also cut its release interval by more than half, reducing the time between Claude Opus 4 and Claude Opus 4.5 from more than six months to around 10 weeks.
Industry experts attribute the global AI companies’ race to release new models faster to price competition. Rather than spending extensive time and enormous sums developing a single, massive top-performing model, companies are shifting their focus toward competing on speed and cost by supplying large numbers of fast, inexpensive models that perform specific tasks well and using them to increase market share.
[email protected] Lee Gu-soon Reporter