BLUN Prince

Fast AI for everyday use

Prince is the fastest and most economical text model in the BLUN family. It was developed for situations where a great many tasks have to be handled without noticeable waiting. Despite its compact size, Prince is a fully capable AI model: it understands natural language, answers questions, processes documents, produces structured data and can write simple programs or automations.

Prince is therefore far more than a model for sorting and forwarding. It can serve as a fast chat assistant, a support model, a data processor or a background model inside digital products. Its strength lies in clearly worded tasks where a direct, dependable answer matters more than a long or deep analysis.

With a context window of up to 16,000 tokens, Prince has room enough for conversations, individual documents, support histories and compact assignments. At the same time it responds considerably faster than the larger models, which makes it well suited to applications with high request volumes.

Context window by tier

• Prince on the free tier: up to 16,000 tokens
• Prince on a paid tier: up to 16,000 tokens

Prince keeps its full context window on both kinds of tier. The model is deliberately optimised for short to medium conversations, individual documents and clearly defined tasks. A paid tier therefore does not inflate the context artificially; above all it makes Prince available for more intensive and more regular use.

A fully capable model – simply faster

With Prince, small does not mean unusable or superficial. The model can grasp how things connect, follow instructions and process information sensibly. It is merely optimised for efficiency and spends less computing power on especially long chains of thought.

Among other things, Prince can:

• answer questions in natural language
• summarise and rephrase content
• extract information from texts and forms
• analyse short documents
• produce structured JSON data
• write simple code components and scripts
• classify requests by topic, urgency or intent
• run standard processes using tools
• handle multilingual content

That makes Prince suitable for many tasks previously handled by rigid rules, simple search systems or manually maintained decision trees.

Ideal for high request volumes

Support systems, online shops, platforms and internal applications generate large numbers of small tasks every day. Not every one of them should be sent to the largest and most expensive model. Prince handles that volume quickly and efficiently.

Typical examples are:

• sorting incoming messages by topic
• answering common customer questions directly
• checking order and form data
• reading out names, amounts, dates or product numbers
• prioritising tickets and routing them to the right team
• tagging and categorising content
• producing short summaries
• preparing input for models further down the line

Where Prince recognises that a task is too complex, too unclear or too risky, the process can pass it on to Queen or King automatically. Maximum computing power is therefore not spent on every request, without underestimating the difficult cases.

Fast chat and digital products

Prince suits applications where an immediate response is what counts. That includes chat windows on websites, assistants inside a piece of software, product help, internal search and short queries in the middle of ongoing work.

Because answers arrive quickly, the interaction feels more natural. Users do not have to wait for a large reasoning model when all they want is a piece of information, an explanation of a setting or a short task carried out.

Automation and structured data

Prince can turn unstructured input into dependable structure. From an email, for instance, it can extract the customer number, the issue, the urgency and the action requested. That information can then be passed straight to a CRM, a ticketing system or an agent.

This makes the model particularly well suited to being the invisible intelligence in the background:

• intent recognition for agents and chatbots
• pre-processing of data
• routing between models and tools
• form and document analysis
• automatic tagging
• checking simple rules
• producing standardised replies
• preparing larger tasks for Queen or King

Where Prince sits in the BLUN model family

Prince is the fast, economical capability tier within the BLUN model family. The model is optimised for short response times, low cost and large volumes of requests, without giving up genuine language understanding and basic reasoning.

Prince is therefore not a replacement for King or Queen on highly complex work. It is the better choice where their additional thinking would bring no measurable benefit.

Typical tasks for Prince

• “Assign this message to a support area and give me the urgency.”
• “Read the invoice number, date and amount out of this text.”
• “Answer this common customer question briefly and clearly.”
• “Summarise this conversation in five bullet points.”
• “Turn these details into a valid JSON object.”
• “Work out which tool this request needs.”
• “Write a small script that renames this file.”

When Prince is the right choice

Use Prince when:

• answers should appear particularly quickly,
• very large numbers of requests are being handled,
• the tasks are short and clearly defined,
• cost per request has to stay low,
• data needs extracting, sorting or structuring,
• a fast model is needed inside an application,
• complex cases can be passed to a larger model when required.

When another model makes more sense

Queen should be used where a task calls for a thorough professional analysis, more extensive code or especially well-crafted wording. King makes sense where many steps depend on one another, where large amounts of context have to be processed, or where long-running agents need directing.