proprietary model

A model whose weightsweights, n.: The learned numbers inside a neural network that set how strongly each part influences the next. The finished set of weights is, in effect, the trained model itself. — and usually its training data, code and architectural details — are kept confidential by its developer and made available only as a service, through consumer applications, a model APImodel API, n.: A web service through which programs send requests to an AI model run by a company and receive its answers, usually paying per use instead of running the model themselves. or managed cloud offerings, under the provider's terms of use. Most 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. have been proprietary, including 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.'s GPT series, 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.'s ClaudeClaude, prop. n.: A family of AI assistants and language models made by the company Anthropic since 2023, widely used for writing, analysis and computer programming. and Google's GeminiGemini, prop. n.: Google's family of AI models and its chatbot assistant, launched in December 2023. Gemini models handle text, images, audio and video and power many Google products.. For developers, such models offer top capability without hardware or maintenance, but bring vendor lock-in, per-tokentoken, n.: The small chunk of text an AI language model reads and writes — often a word, part of a word or a punctuation mark. AI usage, limits and prices are counted in tokens. costs, data-governance concerns, rate limitsrate limit, n.: A cap on how much a customer may use an online AI service in a given time — so many requests or words per minute — beyond which extra requests are refused until later. and model versions that change or are retired on the provider's schedule. For providers, closed release protects commercial advantage and allows usage monitoring, staged access and safety mitigations that cannot be enforced once weights are public. Researchers note that closed models are harder to audit independently and that claims about them are difficult to verify.

We use a proprietary model for the hardest queries and an open-weight one for everything else.