Alphabet Inc. Class C (GOOGL.US) releases Gemini 4 Argon, leading competitors in 13 out of 19 tests; BNP Paribas maintains a $420 price target.

Alphabet Inc. Class C parent company Alphabet (GOOGL.US, GOOG.US) officially unveiled its next-generation frontier AI model Gemini 4 Argon on Wednesday, demonstrating strong competitiveness across multiple areas including complex workflows, knowledge-based tasks, cybersecurity defense, multimodal understanding, and software engineering. The model outperformed competing products from OpenAI and Anthropic in 13 out of 19 benchmark tests, drawing Wall Street's attention. BNP Paribas believes the launch of the new model once again proves Alphabet Inc. Class C's commitment to maintaining competitiveness in the frontier AI space, and maintains an "Outperform" rating on Alphabet with a $420 price target. However, Gemini 4 Argon is not yet fully open to the public. Alphabet Inc. Class C stated that for safety reasons, the model is currently primarily available to the U.S. government and select trusted cybersecurity agencies, with broader access to be gradually expanded after safety protections are further refined. Gemini 4 Argon Delivers Strong Performance, Leading in 13 of 19 Tests As Alphabet Inc. Class C's latest frontier AI model, Gemini 4 Argon has achieved leading performance in multiple key capability tests. According to data released by Alphabet Inc. Class C, Argon surpassed OpenAI's Astra as well as Anthropic's Fable 5.1 and Opus 5.5 in 13 of 19 benchmark tests, demonstrating competitiveness in complex task processing and professional applications. Among these, Argon performed notably well in areas including complex workflows, knowledge-based work, cybersecurity defense, multimodal understanding, and software engineering. Alphabet Inc. Class C also stated that Argon has already been widely used internally, primarily for professional programming tasks, deep research, and writing quality optimization. On commercial pricing, Alphabet Inc. Class C announced that Argon's future usage price will be $2 per million input tokens and $10 per million output tokens. For cached input tokens, Alphabet Inc. Class C will offer a discount equivalent to 95% off the standard input price. This means the effective price for cached input tokens will drop to $0.10 per million, helping to reduce inference costs for repeatedly using contextual information. By offering different pricing and performance options, Alphabet Inc. Class C is expected to further expand the application scope of the Gemini model among developers and enterprise customers. Safety Review Delays Full Release, Alphabet Inc. Class C Emphasizes Rapid Iteration Despite Gemini 4 Argon's strong performance in benchmark tests, Alphabet Inc. Class C has adopted a relatively cautious release strategy. Alphabet Inc. Class C CEO Sundar Pichai said on social media that Argon has been equipped with frontier AI safety protections and is currently being made available to the U.S. government and select trusted cybersecurity defense agencies participating through the Fairwind program. Pichai stated that Alphabet Inc. Class C will expand model access as quickly as possible while maintaining safety, and emphasized that the company will continue rapid iteration to deliver more new developments. Alphabet Inc. Class C DeepMind and Chief AI Architect Koray Kavukcuoglu also stated that the company will continue collecting feedback from early test users to further refine the model's safety protection mechanisms, striving to open Argon to developers, enterprises, and consumers as soon as possible. It is worth noting that this model was originally planned for release around June of this year, when the market widely expected it to be named Gemini 3.5 Pro. However, due to development delays, Alphabet Inc. Class C ultimately adjusted its product roadmap and shifted R&D focus to Gemini 4 Argon. The delay was reportedly mainly related to stricter safety protection requirements and Alphabet Inc. Class C's desire to further enhance the model's programming capabilities before official release. This delay had also raised investor concerns about Alphabet Inc. Class C's AI competitive position. Since the model's delay, Alphabet's market capitalization has shrunk by nearly 20%, reflecting market doubts about whether it can continue to compete with rivals such as OpenAI and Anthropic. BNP Paribas Bullish on Alphabet Inc. Class C's AI Advantages, Maintains $420 Price Target Gemini 4 Argon's debut has received positive reviews from Wall Street analysts. BNP Paribas analyst Nick Jones said in a note to clients that Alphabet Inc. Class C's continued advancement of frontier AI model development is encouraging, especially with Pichai explicitly stating that the company will accelerate model iteration, further boosting market confidence in its AI strategy execution capabilities. Jones believes that Alphabet Inc. Class C holds advantages on multiple fronts in the frontier AI competition, including leading AI research capabilities, proprietary computing infrastructure, extensive product distribution channels, and strong financial resources. These advantages enable Alphabet Inc. Class C not only to continue investing in large model R&D, but also to rapidly extend AI technology to a large number of consumers and enterprise customers by leveraging its existing product ecosystem. Jones expects Alphabet Inc. Class C to continue launching more innovative products around Gemini 4 in the future, including models targeting different application scenarios and offering different pricing options to meet customers' differentiated performance and cost needs. Based on the above assessment, Jones maintains an "Outperform" rating on Alphabet with a $420 price target. Wall Street Awaits Alphabet Inc. Class C's Consumer-Grade AI Agent Launch Beyond model performance improvements, Wall Street is also watching whether Alphabet Inc. Class C will further enter the consumer-grade AI agent market. Jones said it is possible that Alphabet Inc. Class C may launch a consumer-grade AI agent similar to Meta's (META.US) Muse in the future, leveraging Gemini technology and its extensive consumer product ecosystem to drive adoption. Unlike traditional chat Siasun Robot&Automation that mainly provide Q&A and content generation services, AI agents can execute more complex tasks based on user instructions, including information retrieval, workflow planning, and calling external tools to complete operations. For Alphabet Inc. Class C, its massive consumer product ecosystem could provide an important application foundation for AI agents, further extending Gemini's use cases into daily work and life. However, Alphabet Inc. Class C has not yet officially announced the launch of a consumer-grade agent product similar to Muse, and the related assessment remains an analyst expectation regarding the company's future product direction.