Google is preparing to launch a new AI model with stronger coding performance, internally known as "Skimaki," according to a report from the Wall Street Journal cited by India Today. The model, expected to be called Gemini 3.8 Flash, could be released as early as Wednesday, September 2, 2026. If launched as planned, it would represent a significant step in Google's efforts to regain ground in the competitive AI landscape, particularly in the coding segment—a key business use case where Anthropic and OpenAI have taken the lead.
The upcoming release comes at a critical juncture for Google, which has encountered a series of setbacks this year. The company has faced delays in releasing a new flagship model from its most powerful "Pro" series, despite CEO Sundar Pichai stating in May that one would arrive "next month," as reported by Livemint. Internal candidates for the 3.5 Pro model were scrapped because they were not sufficiently better than the Flash series, according to people familiar with the matter cited by Livemint. Meanwhile, Gemini 4, the company's next planned flagship model, has performed well on pretraining evaluations but still needs to complete posttraining, the same source said.
Leadership Changes and Strategic Shift
The launch also comes amid notable leadership changes at Google DeepMind. Co-founder Demis Hassabis stepped aside in an executive shake-up last month, with Koray Kavukcuoglu taking over his day-to-day responsibilities, as reported by both India Today and Livemint. Kavukcuoglu has emphasized to employees his desire to increase the pace of execution, according to Livemint. Reports indicate that Kavukcuoglu had already been leading day-to-day Gemini decisions since at least last year, with Hassabis spending much of his time on external engagements, as noted by India Today.
The leadership transition follows the departure of several high-profile researchers earlier this summer, including Noam Shazeer and Jeff Dean, both cited by India Today and Livemint. These exits added to pressure on Google to accelerate its AI development.
The Flash Model Advantage
The Gemini 3.8 Flash model is part of Google's Flash series, which is designed to be smaller, cheaper, and faster to run compared to the most advanced models from competitors. According to India Today, Flash models are typically trained on a couple hundred billion parameters, whereas the most advanced models from OpenAI or Anthropic use trillions of parameters. This smaller size offers a strategic advantage: it requires less computing capacity to modify, allowing multiple Google teams to test different approaches simultaneously, as reported by Livemint. In contrast, changes to larger Pro models require significantly more of Google's computing resources.
In internal head-to-head tests within Jetski, Google's coding platform, some engineers preferred 3.8 Flash to Anthropic's Opus model, as per the Wall Street Journal report cited by India Today. This suggests that despite its smaller size, the model could deliver competitive coding performance.
Broader Competition and Industry Moves
Google's push to strengthen its coding capabilities is part of a broader industry trend. Since the beginning of the year, Google has redirected more of its researchers' time and computing resources toward improving coding abilities, as reported by multiple sources. The company has also recently hired Barret Zoph, a former Thinking Machines Lab co-founder and OpenAI posttraining lead, as vice president of research focused on reinforcement learning and posttraining, according to both India Today and Livemint.
The competitive landscape has intensified with other tech giants releasing coding-focused models. Elon Musk's SpaceXAI has released new Grok AI models designed for coding, and Meta has launched its own AI coding agent called Muse Code, developed by Alexandr Wang's Superintelligence Labs team, as reported by India Today.
User Adoption and Future Outlook
Despite the setbacks, Google has seen strong user adoption for Gemini, with the app crossing 1 billion users globally, according to India Today. The company co-founder Sergey Brin urged staff to move faster on Gemini after Anthropic released Claude Mythos in April, as reported by the same outlet.
The release of Gemini 3.8 Flash could provide a much-needed boost for Google, though experts caution that a Flash model alone may not fully re-establish Google's position on the AI frontier. As reported by Livemint, a strong debut by Flash 3.8 wouldn't by itself likely be seen as re-establishing Google's place on the so-called frontier. However, the model's performance in coding benchmarks and internal tests could signal progress toward closing the gap with rivals.
With Gemini 4 still in development and the new Flash model poised for release, Google appears to be focusing on immediate gains in coding while working toward its next flagship. The AI race remains highly competitive, and the coming months will be crucial in determining whether Google can regain its edge.