Google rolls out Gemini Deep Think AI, a reasoning model that tests multiple ideas in parallel

Google DeepMind is rolling out Gemini 2.5 Deep Think, its most advanced multi-agent AI reasoning model, to subscribers of its $250-per-month Ultra plan. This model, capable of exploring multiple ideas simultaneously, demonstrates state-of-the-art performance on key benchmarks like Humanity’s Last Exam and LiveCodeBench6, outperforming rival models from xAI and OpenAI. The launch underscores Google's strategic focus on premium, computationally intensive AI offerings and highlights a broader industry convergence on multi-agent systems, signaling a new frontier for competitive differentiation and monetization in the AI sector.

Analysis

Alphabet's Google DeepMind is launching Gemini 2.5 Deep Think, a premium multi-agent AI model, signaling a strategic push into the high-end of the AI market. This launch is notable for its monetization strategy, making the advanced model available exclusively to subscribers of its $250-per-month Ultra plan. The model's technological superiority is substantiated by state-of-the-art performance on key industry benchmarks; it scored 34.8% on Humanity’s Last Exam, significantly outperforming xAI’s Grok 4 (25.4%) and OpenAI’s o3 (20.3%), and also achieved a top score of 87.6% on the LiveCodeBench6 coding test. The move reflects a broader industry convergence towards computationally intensive but more capable multi-agent systems, with competitors like xAI, OpenAI, and Anthropic also developing similar technologies. By gating its most powerful model behind a premium paywall, Google is establishing a clear pathway to directly monetize its AI research and development, aiming to offset the inherently higher costs of multi-agent systems while creating a new revenue tier for enterprise and power users.

Key Decisions for Investors

• The demonstrated performance leadership of Gemini 2.5 Deep Think strengthens Alphabet's competitive positioning in the AI sector, providing a tangible justification for its significant R&D expenditures.
• Investors should monitor the adoption rates of the $250/month Ultra subscription as a key indicator of Google's ability to successfully monetize its most advanced AI capabilities and establish a high-margin revenue stream.
• While the launch signals a current technological lead, the industry-wide shift to multi-agent systems necessitates close tracking of competitor releases from OpenAI, xAI, and Anthropic to assess the durability of Google's advantage.