Alphabet Inc. Class C (GOOGL.US) has launched a new version of Gemini Flash, but the flagship model 3.5 Pro is facing difficulties in production, intensifying concerns over its competitive capabilities at the forefront.

Alphabet Inc. Class C (GOOGL.US) has launched yet another version of its flagship Gemini Flash artificial intelligence (AI) model but has not disclosed when the more powerful AI model, Gemini 3.5 Pro, which is currently facing delays, will be released. In a blog post on Thursday, Alphabet Inc. Class C stated that the newly launched Gemini 3.7 Flash outperforms its predecessor in programming tasks such as debugging and has a stronger capability to produce deployable, production-ready code right from the first generation. The company added that the new model can complete application development with fewer prompts, offering a better developer experience and reducing the cost of tokens required to run the model. The AI productivity agent Gemini Spark from Alphabet Inc. Class C will be supported by 3.7 Flash starting Thursday. Alphabet Inc. Class C further stated that to fulfill its commitment to "Frontier Safety," the company has enhanced the security features of Gemini 3.7 Flash, including measures to block malicious hacking attempts and prevent the misuse of hazardous chemical, biological, radioactive, or nuclear materials, all while ensuring that safe and beneficial operations are not hindered. Despite the ongoing expansion of the Gemini Flash product line, Alphabet Inc. Class C continues to struggle to keep pace with competitors OpenAI and Anthropic in the high-stakes competition to build cutting-edge AI models. The delay of Gemini 3.5 Pro has led investors to start questioning the product roadmap of Alphabet Inc. Class C, especially in areas like AI programming that hold significant commercial value. At the I/O Developer Conference in May this year, Sundar Pichai stated that the Gemini 3.5 Pro model would be released in June, but it has yet to go online. Originally, the Gemini 3.5 Pro model was supposed to take on the task of re-establishing Alphabet Inc. Class C's competitive edge in the AI frontier. After all, against the backdrop of OpenAI and Anthropic consistently enhancing their model capabilities, the Gemini Pro series has long been a key benchmark for assessing the AI strength of Alphabet Inc. Class C. However, the inability to determine the release date of the next flagship model has further fueled widespread doubts about whether this tech giant can surpass its competitors and successfully transform its substantial AI investments into leading tools and services in the market. In a financial conference call held by Alphabet Inc. Class C in July, CEO Sundar Pichai stated that the company plans to launch models at a faster pace and mentioned that it has invested significant computing resources to train the upcoming Gemini 4 model. Just before the earnings release, this tech giant also launched three Flash versions aimed at achieving higher efficiency and quality. There is speculation that the actual capabilities of the Gemini 3.5 Pro model may not meet the initial expectations of Alphabet Inc. Class C. The industry analysis firm Semi Analysis believes that the capabilities of Gemini 3.5 Pro are roughly comparable to those of Anthropic Claude Opus 4.5, which was released at the end of November last year. The firm even indicated in its latest report that the Gemini 3.5 Pro model may have already been shelved internally at Alphabet Inc. Class C. According to insiders, Alphabet Inc. Class C has been taking time to improve the capabilities of Gemini 3.5 Pro, particularly in programming, which has resulted in the model's launch being delayed by several months. Ten current and former employees revealed that this delay has frustrated engineers, AI researchers, and management at Alphabet Inc. Class C, many of whom are concerned that as Anthropic and OpenAI continue to launch models that surpass Gemini's capabilities, Alphabet Inc. Class C may lose its competitive edge in the market. Insiders noted that the process of preparing the model for release involves multiple layers of stakeholders, while also striving to integrate AI into a vast product ecosystem that includes Search, Maps, and YouTube, which may contribute to delays in the release process. With top talents leaving one after another and a major shuffle in the AI leadership, Alphabet Inc. Class C has recently seen some turbulence. In addition to the delay of the Gemini 3.5 Pro model, the recent talent exodus at Alphabet Inc. Class C is also concerning. Earlier this month, Chief Scientist Jeff Dean announced his departure after 27 years. Several renowned researchers have previously left Alphabet Inc. Class C, including Noam Shazeer, one of the authors of the landmark 2017 paper "Attention Is All You Need," which laid the foundation for generative AI; now all eight authors have left the company. Shazeer transitioned to OpenAI in June, less than two years after Alphabet Inc. Class C recruited him back through acqui-hiring for nearly $3 billion. Shortly after his departure, Nobel laureate John Jumper also left DeepMind to join Anthropic. D.A. Davidson analyst Gil Luria pointed out a noticeable trend of top talent leaving Alphabet Inc. Class C. He stated, They are not enthusiastic about commercializing AI; they want to be part of history. Therefore, they see Anthropic, OpenAI, or other startups as places where they can write history. Alphabet Inc. Class C's investment scale in global data centers, chips, and related infrastructure is nearly unmatched, but computing power remains scarce. Each TPU allocated for training models, supporting Alphabet Inc. Class C's products, or fulfilling cloud customer contracts represents a choice among multiple priorities. According to multiple anonymous insiders, some researchers at Alphabet Inc. Class C are increasingly dissatisfied with the access to computing powerthey find it difficult to obtain the computational resources necessary to advance cutting-edge projects but are seeing Alphabet Inc. Class C sell self-developed TPUs to external customers, including Anthropic. Furthermore, the hierarchical approval process within Alphabet Inc. Class C is cumbersome, and the transformation of research achievements into products requires passing through multiple levels, making OpenAI, Anthropic, and even younger startups more attractive to AI developers who prefer lab work over financial report figures. However, Alphabet Inc. Class C's AI leadership has undergone a significant restructuring, seemingly indicating that the company is preparing to face challenges. Reports indicate that DeepMind co-founder Demis Hassabis has handed over day-to-day management duties and stepped down as the overall leader responsible for commercializing Gemini, transitioning to the role of chairman at DeepMind as well as Alphabet Inc. Class Cs Chief Scientist, focusing primarily on long-term AI research. Former DeepMind Chief Technology Officer Koray Kavukcuoglu has been promoted to Senior Vice President of DeepMind (DeepMind will no longer have an independent CEO) and will be responsible for the development and operation of the Gemini model, directly reporting to Pichai. Additionally, Alphabet Inc. Class C co-founder Sergey Brin will be more directly involved with Gemini. Currently, Brin does not hold an official executive position. However, since the release of ChatGPT, he has re-engaged in Alphabet Inc. Class Cs daily AI affairs, participating in model testing, discussing technical directions, and exerting increasing influence over the development direction of Gemini. The core of this reorganization is to move the AI decision-making center from London back to Silicon Valley, aiming to reverse a trend that has plagued the company at least since 2023. That year, Alphabet Inc. Class C merged two previously independent and highly valued research laboratories: Google Brain, based at the companys Mountain View headquarters, and DeepMind, rooted in London. Although the two labs merged under the name Alphabet Inc. Class C DeepMind, researchers remained working on different continents. According to insiders, this arrangement complicated the decision-making process and left talents in both locations dissatisfied. A key focus of Alphabet Inc. Class Cs future business strategy is to transform the research federation, previously scattered between Londons DeepMind and Californias Google Brain, into a Gemini product delivery machine centered in Mountain View and directly accountable to Pichai. Kavukcuoglu holding the positions of daily operations of Google DeepMind and Chief AI Architect means that model pre-training, post-training, evaluation, computing power scheduling, and coordination with Cloud, Search, and developer products will be integrated into a shorter decision-making chain. These latest personnel adjustments are not intended to weaken fundamental research but rather to separate the management of long-term scientific exploration and quarterly product delivery, in an attempt to address the slow decision-making and commercialization issues affecting the post-merger cross-continental teams since 2023. The Gemini 3 series models and Nano Banana image editing tool have helped Alphabet Inc. Class C regain market attention. Alphabet Inc. Class C has also demonstrated that it still possesses the capability to develop cutting-edge models. However, the more pressing task is to rapidly convert model capabilities into developer tools, enterprise services, and stable revenue. In the fields of AI programming and the enterprise market, OpenAI and Anthropic have already established a first-mover advantage. Executives and the board at Alphabet Inc. Class C are concerned that the companys research strength has not yet fully translated into product competitiveness. Previously, Gemini simultaneously hosted functions including research, modeling, application, and security. The long decision chain often led to conflicts between research objectives and product rhythms. The new management structure assigns daily development of Gemini to Kavukcuoglu, which should help streamline decision-making cycles and accelerate model releases and product rollouts. Reports also mention that some executives at Alphabet Inc. Class C are dissatisfied with Hassabis's level of investment in commercialization efforts. AlphaFold is considered one of the controversial cases. This protein structure prediction system helped Hassabis win the 2024 Nobel Prize in Chemistry and has brought enormous impact on scientific research, but after the project was opened for free access, Alphabet Inc. Class C did not achieve commercial returns commensurate with the scale of investment. Those close to Alphabet Inc. Class C deny any significant conflicts between the two parties. They noted that Hassabis led the early release of Gemini and has always been concerned about the project's progress. Whether the new structure at Alphabet Inc. Class C will prove effective following the significant changes in the AI leadership depends on whether three paths can align smoothly: research needs to retain exploration space, model teams need to increase iteration speed, and product departments must identify users willing to pay. For Alphabet Inc. Class C, top-tier papers and model rankings are no longer sufficient to dispel external doubts; the ensuing competition will focus on code tools, enterprise markets, product experience, and commercial revenue. Brin's return to the center of AI power indicates that Alphabet Inc. Class C has regarded Gemini as a core battle that requires the involvement of the founders. Hassabis stepping back to frontline research also means that the research-driven model established over the past decade at DeepMind is undergoing change. Alphabet Inc. Class C possesses models, computing power, data, distribution channels, and a vast enterprise customer base. Now, it needs to prove it can compress these resources into a faster product delivery pipeline.