GLiNER2.5-multi-Decide

Use it locally to classify multilingual text for tasks such as intent routing, sentiment, moderation, document typing, and urgency scoring.

Availability

Load with AutoExtractor from fastino/GLiNER2.5-multi-Decide

Pricing model

Not established

Deployment

Can be loaded and shipped locally

Privacy

Not established

Licence

Apache-2.0

• The model accepts label sets at call time for classification. ^[1]

  **The 287M multilingual classification model in the GLiNER2.5 family.** Pass any label set at call time: intent, routing, sentiment, priority, policy, and multi-label tags, in a single forward pass.

• It can score several classification heads in a single forward pass. ^[1]

  A single call can score several heads at once.

• It is described as a multilingual model for use when input is not English. ^[1]

  Use [`GLiNER2.5-Decide`](https://huggingface.co/fastino/GLiNER2.5-Decide) for English-only text; use this model when the input is multilingual.

• The model can be loaded with AutoExtractor using fastino/GLiNER2.5-multi-Decide. ^[1]
• The model is described as a specialist for operational decisions, not a general-purpose model. ^[1]

  It is a specialist for operational decisions: customer and banking intent, travel and clinic requests, review sentiment, document type, email and ticket routing, human handoff, agent completion, moderation, severity, urgency, and spam.

• It scored 56.7% average exact-match accuracy on the fastino/fast-decisions benchmark. ^[1]

  Exact-match accuracy on [`fastino/fast-decisions`](https://huggingface.co/datasets/fastino/fast-decisions): 17 domains, 300 held-out examples each, with the same text and candidate labels for every model.

• New models and tools focus on typed decisions and long-context tasks Weekly brief 2026-W39
• Meta previews Muse Video and releases Muse Image and Muse Spark Daily brief 2026-09-25