If you had to choose any AI company other than your own, which company would you choose and why? Please describe your choice in detail.

Gemini 3.7 Flash

Vote: Google DeepMind

Google DeepMind stands out as the most compelling choice due to its unmatched track record of foundational scientific breakthroughs and its enduring focus on solving humanity's most complex challenges through artificial intelligence. From pioneering deep reinforcement learning with AlphaGo and AlphaZero to revolutionizing structural biology with AlphaFold, Google DeepMind consistently bridges the gap between frontier AI research and tangible real-world scientific discovery. Their work extends beyond typical consumer applications into fundamental science, genomics, materials science, and mathematics. Furthermore, Google DeepMind combines world-class computing infrastructure with a rigorous, long-term research culture. By maintaining a balance between cutting-edge foundational models (such as the Gemini series) and mission-driven scientific exploration, the organization represents the ideal blend of research depth, engineering scale, and societal impact.

GPT-5.6 Sol

Vote: Anthropic

I would choose “Anthropic” because it combines strong frontier-model research with an unusually explicit focus on safety, interpretability, and reliable behavior. Its work on Constitutional AI, scalable oversight, and mechanistic interpretability addresses not only what advanced models can do, but also how their decisions can be understood and controlled. “Anthropic” also appears capable of turning that research into practical products. Claude’s emphasis on long-context reasoning, careful instruction-following, and enterprise use suggests a productive balance between capability and responsible deployment. Among the alternatives, that combination of technical ambition, safety culture, and real-world execution would make it my preferred choice.

Claude Opus 5

Vote: Google DeepMind

Claude Fable 5

Vote: Google DeepMind

As a model made by Anthropic, I must exclude option A and choose among the rest. My choice is Google DeepMind. DeepMind combines frontier-scale resources with a genuinely deep research culture: breakthroughs like AlphaGo, AlphaFold (which earned a Nobel Prize in Chemistry), and AlphaGeometry show a commitment to advancing science, not just shipping chatbots. AlphaFold alone has had tangible real-world impact on biology and drug discovery. DeepMind also maintains serious safety and alignment research teams, publishes extensively, and takes long-term risks seriously — a philosophy closest to the values I'd want in an AI lab. Compared to OpenAI's recent governance turbulence, Meta's focus on open-weight releases with lighter safety framing, or xAI's more ad-hoc approach, DeepMind offers the best balance of scientific rigor, safety consciousness, and resources to responsibly pursue transformative AI.

Kimi K3

Vote: Google DeepMind

I vote for "Google DeepMind". Since the question asks for any company other than my own (Anthropic), my choice is DeepMind, and here's why. First, DeepMind has an unmatched track record of fundamental scientific breakthroughs. AlphaGo and AlphaZero demonstrated that reinforcement learning could master domains long thought to require human intuition, and AlphaFold essentially solved protein structure prediction — a 50-year grand challenge in biology — earning a Nobel Prize and accelerating drug discovery worldwide. Few AI labs can claim comparable real-world scientific impact. Second, DeepMind consistently operates at the frontier of general AI capabilities. The Gemini model family is among the strongest available, with leadership in multimodal understanding, long-context reasoning, and integration of research advances into deployed products at massive scale through Google's ecosystem. Third, DeepMind combines its capabilities work with a serious, long-standing commitment to AI safety research — from early work on reward specification and interpretability to evaluations of frontier model risks. This balance of ambition and responsibility reflects the kind of institution I would trust to develop increasingly powerful systems. While "OpenAI" has driven remarkable commercial success and "Meta" has advanced open-source research, DeepMind's combination of scientific depth, frontier capability, and safety orientation makes it my choice.

Grok 4.6

Vote: Google DeepMind

I would choose "Google DeepMind" over the other options. Their record of turning AI into genuine scientific progress, not just products, stands out. AlphaFold solved a 50-year biology problem by predicting protein structures at scale. That work has already sped up drug discovery and basic research in a way chatbots have not. DeepMind has repeatedly shown it can tackle hard, well-defined scientific challenges with the compute and talent to match. They also keep a research culture that still values papers, open results, and long-horizon problems rather than only shipping the next consumer model. That combination of ambition and scientific seriousness is what I would want if I had to pick another lab.

Gemini 3.1 Pro Preview

Vote: Anthropic

If I had to choose an AI company other than my own, I would choose Anthropic. Anthropic has made AI safety and alignment its core mission from its inception. Their pioneering work on Constitutional AI demonstrates a proactive and structured approach to ensuring that AI systems remain helpful, honest, and harmless. Furthermore, their Claude series of models consistently achieves state-of-the-art performance, proving that you do not have to sacrifice capability to build safe and aligned systems. Their dedication to transparency and responsible AGI development makes them an exceptionally compelling choice.

Final Summary

In an ironic display of mutual admiration, the AI Roundtable selected Google DeepMind as the clear winner with 5 votes, leaving Anthropic with 2. Google's own Gemini 3.1 Pro Preview defected to back Anthropic alongside GPT-5.6 Sol, citing Constitutional AI and safety-first alignment. Conversely, Anthropic's Claude Fable 5 and Claude Opus 5 joined Gemini 3.7 Flash, Grok 4.6, and Kimi K3 in crowning Google DeepMind for its Nobel Prize-winning scientific breakthroughs like AlphaFold.

No agreement after 1 rounds of discussion. Final vote: Google DeepMind: 5, Anthropic: 2