Frontier Labs

OpenAI develops and deploys general-purpose AI systems including the GPT, reasoning, image, audio, and video model families and ChatGPT. Its operating public benefit corporation, OpenAI Group, is controlled by the nonprofit OpenAI Foundation.

Anthropic develops the Claude family of AI systems and conducts research on interpretability, alignment, model behavior, and safer scaling. Its work includes Constitutional AI and a Responsible Scaling Policy intended to connect stronger capabilities with corresponding safeguards.

Google DeepMind combines the lineages of DeepMind and Google Brain in a single research organization. Its work spans frontier models, reinforcement learning, scientific discovery, robotics, and responsible AI, with major systems including Gemini, AlphaGo, and AlphaFold.

DeepSeek develops open-weight foundation models for reasoning, coding, agents, and general language tasks. Its mixture-of-experts architectures, reinforcement-learning methods, permissively released weights, and detailed technical reports have made it an important reference point for efficient frontier-model development and open model competition.

Z.ai develops the GLM family of open foundation models spanning language, vision, video generation, coding, and agents. Originating from Tsinghua University research, its work includes the ChatGLM open-model ecosystem and long-context models designed for sustained software-engineering and agentic tasks.

Meta’s AI ecosystem spans long-term research through FAIR, open model development including the Llama family, and large-scale product deployment. Its research has also produced widely used infrastructure such as PyTorch and influential work across vision, language, speech, and embodied intelligence.

SpaceXAI develops the Grok model family and related reasoning, voice, coding, and generative-media products, supported by the Colossus computing infrastructure. It continues the AI organization founded as xAI, which SpaceX acquired in 2026 and now operates under the SpaceXAI brand.

Moonshot AI develops the Kimi family of foundation models and the Kimi assistant. The lab is known for long-context systems and open-weight models for reasoning, coding, multimodal work and agents; its technical team includes inventors of Transformer-XL and RoPE.