frontier lab
In plain wordsOne of the handful of companies building the most advanced AI models of the moment, spending billions of dollars on computing power to train them.
An organization that develops frontier modelsfrontier model, n.: One of the most capable AI models in existence at a given time, built by a handful of leading labs at great cost, and the main target of AI safety rules., the most capable general-purpose models at a given time, typically with very large compute budgetscompute budget, n.: The total amount of computing a team can spend on training (or running) a model, set in advance like a money budget and used to decide how big the model should be and how much data it sees. and in-house research, training and safety teams. As of October 2026, commonly cited examples include OpenAI, Anthropic, Google DeepMind, Meta and xAI in the United States and DeepSeek and Alibaba's Qwen team in China; membership changes as capabilities shift. Frontier labs are the main addressees of voluntary pledges such as the Frontier AI Safety CommitmentsFrontier AI Safety Commitments, prop. n.: Promises made in 2024 by major AI companies to publish safety plans for their most advanced models, including the danger points at which they would stop if the risks could not be controlled., industry bodies such as the Frontier Model ForumFrontier Model Forum, prop. n.: An industry group set up in 2023 by leading AI companies to share safety research and best practices for the most advanced AI models., company safety frameworks (responsible scaling policiesresponsible scaling policy, n.: A public rulebook in which an AI company promises to test each new model for dangerous abilities and to add stronger safeguards, or slow down, as its models become more capable., the Preparedness FrameworkPreparedness Framework, prop. n.: OpenAI's published process for testing its AI models for catastrophic risks — such as help with biological weapons or cyberattacks — and deciding what safeguards are needed before release., the Frontier Safety FrameworkFrontier Safety Framework, prop. n.: Google DeepMind's published plan for spotting when its AI models approach dangerous abilities and applying stronger security and safeguards before that happens.) and laws aimed at the largest developers, such as California's SB 53SB 53, prop. n.: A 2025 California law requiring developers of the most powerful AI models to publish their safety plans, report serious safety incidents to the state and protect employees who raise alarms. (2025), which places transparency and incident-reporting duties on 'frontier developers' (those that have trained a model using more than 10²⁶ operations) and further duties on those with annual revenue above $500 million.
Only frontier labs can afford a training run of that size, so the paper's results are hard to replicate.
Origin
Derived from 'frontier AI', popularized in 2023 by Markus Anderljung and colleagues' paper 'Frontier AI Regulation' (July 2023), the Frontier Model Forum founded that month by Anthropic, Google, Microsoft and OpenAI, and the UK's Frontier AI Taskforce and AI Safety SummitAI Safety Summit, prop. n.: A series of international meetings of governments, AI companies and researchers, begun in England in 2023, to agree on how to handle the risks and benefits of the most advanced AI. (November 2023). 'Frontier lab' is the informal shorthand; earlier writing said 'AGI labs' or 'leading AI labs'.
Usage note
The label is relative and contested: who counts as frontier shifts with each model release, and critics argue it flatters companies or implies an inevitable race. Legal texts prefer precise thresholds on training compute or revenue.
Related
- frontier modelfrontier model, n.: One of the most capable AI models in existence at a given time, built by a handful of leading labs at great cost, and the main target of AI safety rules.
- frontier AIfrontier AI, n.: The most capable general-purpose AI systems of the moment — those at or beyond the leading edge — which governments single out for extra safety attention.
- Frontier Model ForumFrontier Model Forum, prop. n.: An industry group set up in 2023 by leading AI companies to share safety research and best practices for the most advanced AI models.
- responsible scaling policyresponsible scaling policy, n.: A public rulebook in which an AI company promises to test each new model for dangerous abilities and to add stronger safeguards, or slow down, as its models become more capable.
- compute budgetcompute budget, n.: The total amount of computing a team can spend on training (or running) a model, set in advance like a money budget and used to decide how big the model should be and how much data it sees.
- AI arms raceAI arms race, n.: A competition among countries or companies to build the most powerful AI first, in which fear of falling behind pushes everyone to move faster and cut corners on safety.
- concentration of powerconcentration of power, n.: The worry that AI will put enormous economic, political or military power in the hands of a few companies, governments or individuals, beyond the reach of ordinary checks and balances.
Sources
- Anderljung, M. et al. (2023). Frontier AI Regulation: Managing Emerging Risks to Public Safety. arXiv:2307.03718. arxiv.org/abs/2307.03718
Mentioned in
AI arms raceAI arms race, n.: A competition among countries or companies to build the most powerful AI first, in which fear of falling behind pushes everyone to move faster and cut corners on safety., AnthropicAnthropic, prop. n.: An American AI company that makes the Claude chatbot and models. It was founded in 2021 by former OpenAI staff and puts particular emphasis on AI safety research., frontier AIfrontier AI, n.: The most capable general-purpose AI systems of the moment — those at or beyond the leading edge — which governments single out for extra safety attention., frontier modelfrontier model, n.: One of the most capable AI models in existence at a given time, built by a handful of leading labs at great cost, and the main target of AI safety rules., OpenAIOpenAI, prop. n.: The American company that makes ChatGPT and the GPT models. It began in 2015 as a nonprofit research lab and is now a for-profit company overseen by a nonprofit foundation., superintelligencesuperintelligence, n.: A mind, usually imagined as an AI, that would leave even the cleverest people far behind in nearly every kind of thinking, not just in one game or skill. None is known to exist., Meta Superintelligence LabsMeta Superintelligence Labs, prop. n.: A division Meta created in 2025 to pursue superintelligence, gathering all of the company's AI teams under new leaders recruited in an expensive hiring campaign., moatmoat, n.: A lasting business advantage that keeps competitors out, like a castle's moat; in AI, the question of whether any company can stay ahead when rivals and free models catch up quickly., recursive self-improvementrecursive self-improvement, n.: An AI improving its own design or training, so that each improved version is better at making the next improvement. It is the engine usually imagined behind an intelligence explosion., Thinking Machines LabThinking Machines Lab, prop. n.: Mira Murati's AI company, which she started in San Francisco in 2025 after serving as OpenAI's technology chief. It makes tools for customizing AI models and releases some models openly., Jan LeikeJan Leike, prop. n.: A German AI-alignment researcher who helped develop training from human feedback, co-led OpenAI's superalignment team, and now leads alignment research at Anthropic., reverse acquihirereverse acquihire, n.: A deal in which a big tech company hires a start-up's founders and top staff and pays to license its technology, without buying the company itself.