Simular

Simular builds AI agents that autonomously operate computers and smartphones like humans, achieving super-human performance on benchmarks through research in agentic AI and reinforcement learning.

Founded in 2023 by former DeepMind researchers, Simular is building AI agents that interact with computers and smartphones in the same manner as humans do - a vision that has earned the firm the moniker "the Autonomous Computer Company." The team comprises scientists and engineers drawn from DeepMind, Google, and other leading AI laboratories, and it maintains a research-driven culture that publishes regularly at top conferences while translating findings directly into deployed products.

The company's technical work spans agentic AI, neuro-symbolic frameworks, lifelong reinforcement learning, large language models, and symbolic code, underpinned by contributions to open-source frameworks. Its flagship product, Agent S3, demonstrates the practical payoff of this approach: the system has achieved a 72.6% success rate on the OSWorld benchmark - surpassing human-level performance - alongside 90.1% on the WebVoyager browser benchmark and 71.6% on AndroidWorld smartphone tasks. Over 12,600 users have already joined the platform.

Simular has secured $21.5M in funding to advance its mission. For engineers and researchers drawn to work at the frontier of autonomous AI, where theoretical advances are expected to produce tangible, real-world impact, the company offers a rare convergence of academic rigour and product ambition.