EN | IT
Products Infrastructure
Our Story Team Investors Contact Us
Careers Events News

RESOURCES / NEWS

Introducing: Xference v0.8

12 Oct, 2026 ▪ Cagliari
Xference 0.8 Release

Starting this October, we’re rolling out our new version: Xference 0.8. Here’s what it can do, in a nutshell:

  • Ability to use web search, ask anything, get a clear answer, from everyday questions to science, history or technology — explained simply, at the level you want.
  • Ability to use and read images.
  • Ability to create PDF, DOC, PowerPoint, and XLS files — attach a document, spreadsheet or image and ask it to summarize it, answer questions about it.
  • Ability to select a preferred open-source LLM model
  • Ability to enable or disable "Thinking mode"

Our models

You do not need to be an expert to choose. We run three models, and each is best at a different kind of help. Here is a plain-language guide to when each one is most useful for you.

Qwen 3.8-27B

The Deep Thinker

Think of this one as the patent expert you hand a difficult problem to. Use it when the topic is complex, the material is long, or you want a thorough, careful analysis.

  • Complex or tricky questions that need careful, step-by-step reasoning
  • Looking at images: photos, charts, screenshots, documents
  • Long material: it can work through an entire contract, a long report or a lengthy conversation without losing the thread
  • Multi-step projects: for example, drafting a document, revising it and polishing the final version in one flow

Gemma 4 12B

The Fast Everyday Helper

Think of this one as the quick, nimble assistant for day-to-day moments. Use it when you are working on something short and simple.

  • Quick questions that need a short, fast answer
  • Everyday writing — summaries, translations, brainstorming, drafting messages
  • Using images and/or voice
  • Fast, light replies, responsive and snappy for simple tasks

Qwen3.6-35B-A3B

The Smart All-Rounder

Think of this one as the well-read generalist who is also quick on their feet. It combines the knowledge of a large model with the speed of a small one. Use it when you want a good answer without the wait.

  • A balance of quality and speed: solid answers for everyday work, with reasoning when the question needs it
  • Looking at images: photos, charts, screenshots
  • Long documents and long conversations, such as reports, contracts or extended chats
  • Writing and checking code, working with structured data, and tasks that use tools such as web search

Below is a quick guide when choosing a model based on common tasks.

If you want to... Use
Ask a quick question and move onGemma 4 12B
Go deep on a complex topicQwen3.8-27B
Have a photo or chartQwen3.8-27B
Work through a long documentQwen3.8-27B
Get a fast answer for everyday tasksGemma 4 12B
Get a good balance of speed and qualityQwen3.6-35B-A3B

Choose Xference your fast, flexible, reliable private inference platform.

CONTACT US

EuropeAI Private Inference Technology made in Europe