Gemma 4 26B A4B IT

Google DeepMind

Catalogue only Open weights

Sovereignty and hosting

Catalogue only

This model is catalogued for reference. It is not hosted on LLM EU yet, and no first-party endpoint serves it here. Requests to it go to the provider directly, under that provider's own terms.

Summary

Gemma 4 26B A4B IT is a mid-size Gemma 4 model stored as 25.8B parameters, sitting between the 12B and 31B sizes in the family. It has the same 256K-token context window as the 31B model and takes text and image input. The model card reports 82.6% on MMLU Pro and 88.3% on AIME 2026 without tools.

Specifications

Provider
Google DeepMind (US)
Context window
262k
Parameters
25.81B
Modalities
text, vision
License
apache-2.0 · licence
Open weights
yes
Release date
11 Mar 2026
Tasks
reasoning, coding, vision, summarization, cheap

Languages

Strong: en, de, fr, es, it, pt

Adequate: nl, pl, sv, cs, ro, el, uk, ru, ar, hi, zh, ja, ko, tr

Languages

Strong

  • en
  • de
  • fr
  • es
  • it
  • pt

Adequate

  • nl
  • pl
  • sv
  • cs
  • ro
  • el
  • uk
  • ru
  • ar
  • hi
  • zh
  • ja
  • ko
  • tr

Not verified: the provider has not documented this claim. — a language is listed as strong only where a source claims it; the card links the evidence where one exists.

When to use it

Strengths

  • mid-size option with the 256K-token window of the larger model
  • Apache-2.0 licence
  • text and image input
  • vendor reports near-flagship scores at a smaller footprint

Limits

  • the "A4B" suffix is the vendor’s own designation and is not defined on the model card
  • scores are self-reported
  • no evaluation has been run by LLM EU