Venice Adds Unbiased’s Pareto Model With a One-Million-Token Window

Venice added Unbiased’s Pareto model on Nov. 12, giving users of its anonymous-access AI platform image input and a one-million-token context window for larger requests.

A context window measures how much material a model can consider in one request. Tokens are pieces of text that models read and generate. A window of roughly one million tokens lets users include more material at once, such as a codebase or a stack of reports, and ask questions across it without splitting it into smaller submissions. Pareto also accepts image input, allowing visual material alongside text.

Pareto is a composite system that processes responses from several frontier and open models in parallel, then combines them into one answer. A router directs a request to a single model; Pareto draws on multiple model outputs for each response.

Unbiased first released Pareto on Sept. 17. Its Oct. 1 26.10 Preview update expanded the context window from 262,000 tokens to 1,048,576. 26.10 Preview is priced at about $0.80 per million input tokens and $3.20 per million output tokens. Input tokens are material sent to the model, while output tokens are generated in response, making the listed output rate four times higher. The October update also lowered prices for older Pareto versions.

Venice’s platform prioritizes anonymous access and limited data retention. Users can store conversations in their browser rather than on Venice’s servers. The platform also offers chat, image and video tools, along with API access.

Pareto joins Inception Labs’ Mercury 3, which Venice lists for API use, another model with a one-million-token context window.