DeepMind's SIMA 2: An AI Agent for Video Games and Robotics
DeepMind has unveiled SIMA 2, an agent that has evolved from an obedient executor into an interactive companion. It no longer just follows commands: it reasons about goals, talks with the user and improves itself. The agent sees the screen the way a person does and controls a mouse and keyboard. It has the Gemini model built in. Researchers call it a step toward AGI and the future of personal robots.
AI-processed from DeepMind Blog; edited by Hamidun News
DeepMind has unveiled SIMA 2, a neural network for video games and robotics. It is an agent that has evolved from following simple instructions into an interactive helper that can reason, hold a conversation and improve itself.
How does SIMA 2 differ from the first version?
The first SIMA performed more than 600 skills in video games from instructions ("turn left", "climb the ladder", "open the map"). SIMA 2 has evolved into an interactive helper that can do more than follow commands: it can also reason, hold a conversation and improve itself.
How does the SIMA 2 neural network control video games?
SIMA 2 acts like a human player: it looks at the screen and operates a virtual keyboard and mouse, without access to the game's internal mechanics.
- Nov 2, 2026Bun creator uses Claude Code to port 750K lines from Zig to Rust in 11 days
- Nov 2, 2026The Verge reviewer builds a $50 Meta Muse Charm alternative with Claude Code
- Nov 1, 202616 models tested on 7 search tools: Exa averaged 10.5 out of 14, SearXNG about 3
- Nov 1, 2026Google's StreamVibe video agent beats other models by 10.7% on response timing and quality
Key Questions Answered
How does SIMA 2 differ from the first version?
The first SIMA performed more than 600 skills in video games from instructions ("turn left", "climb the ladder", "open the map"). SIMA 2 has evolved into an interactive helper that can do more than follow commands: it can also reason, hold a conversation and improve itself.
How does the SIMA 2 neural network control video games?
SIMA 2 acts like a human player: it looks at the screen and operates a virtual keyboard and mouse, without access to the game's internal mechanics.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.