Fastino Labs
2026Fastino Labspaid
GLiNER Model Week
Open Source
Small Language Models
Fastino Labs celebrates 30 million GLiNER downloads by releasing two new open-source SLMs pushing the parameter efficiency frontier. One matches SOTA accuracy at up to 93x smaller parameter count; the other beats a recent frontier model on key benchmarks. Both released on Hugging Face.
First new SLM matches SOTA accuracy at up to 93x fewer parameters
Second new SLM outperforms a recent frontier model on key benchmarks
Released open-source on Hugging Face to celebrate 30 million GLiNER downloads
pricingFree (Open Source)
regionUnited States
2026Fastino Labsopen source
GLiNER
v2.5
NLP
Information Extraction
GLiNER2.5 is a significant upgrade to the GLiNER architecture that replaces span enumeration with boundary prediction for more efficient, scalable information extraction. It introduces five key capabilities, including long-context processing, unlimited span length, joint entity-relation extraction, constrained classification, and span-level attributes.
Boundary-prediction architecture for linear inference scaling
Unlimited span length for extracting entities of any size
Joint entity and relation extraction for coherent knowledge graphs
pricingOpen-weight under Apache 2.0 license
version2.5