Languages

Falcon 2 11B Languages: What It Was Trained On

Falcon 2 11B is multilingual, but not evenly. TII lists the languages it was trained on and warns explicitly about the ones it was not.

Back to the overview

The documented list

  • Falcon 2 11B: English, German, Spanish, French, Italian, Portuguese, Polish, Dutch, Romanian and Czech, with Swedish also named in the model detail section.
  • Falcon 2 11B VLM: English only.
  • The model card states the model will not generalise appropriately to languages outside that set.

Why the warning is worth taking literally

Multilingual claims are usually soft, so an explicit limitation on the model card is a gift. It tells you the failure mode is predictable rather than surprising: output in an unsupported language degrades towards English or towards incoherence, and it does so more readily than a model with a wider pretraining mix.

What to check before shipping

  • Test in the exact language and script your users write in, not in English.
  • Check numerals, dates and currency formatting per language.
  • Remember the VLM cannot read non-English questions reliably at all.

Related pages

Training dataFalcon 2 11B VLM

Sources

falcon2.lol is an independent third-party site. It is not affiliated with, endorsed by, sponsored by or operated by the Technology Innovation Institute, and it does not speak for TII. Every number on this page is attributed to the source listed above.

Falcon 2 11B Languages: What It Was Trained On