Fastino has released GLiNER2.5, a named entity recognition (NER) model that replaces span enumeration with boundary prediction [reference:8]. The release removes the maximum entity width limit, supports a 4,096-word context, and keeps computation linear in sequence length for a fixed schema [reference:9]. The model also unlocks joint entity-relation decoding, cross-task label constraints, and per-span attributes.
Three checkpoints under Apache 2.0
Fastino released three GLiNER2.5 checkpoints on Hugging Face under Apache 2.0 at 74M, 194M, and 287M parameters [reference:10]. The models support local inference on CPU, CUDA, or MPS via pip install "gliner2[local]" (Python 3.10+) [reference:11]. No inference provider currently hosts the checkpoints, so self-hosting is the deployment path.
Performance gains
Across 16 zero-shot benchmarks, the multilingual checkpoint reaches 56.17 overall macro F1 against 56.09 for GLiNER2, with a 24.75-point gain on XNLI [reference:12]. The headline gain is XNLI, where Multi jumps to 62.30 from 37.55 [reference:13]. Few-NERD improves for Base to 55.14 from 47.22, and Romanian RONEC, an untrained language, also improves.
Architectural innovation
Earlier GLiNER models located entities by candidate spans: every start position paired with every allowed width, each scored against the schema. That design tied compute to a width axis and imposed a hard ceiling on entity length. GLiNER2.5 removes enumeration. The shared encoder processes text and schema queries in one pass. Instead of scoring spans, the model predicts start and end scores over token boundaries plus inside scores over tokens. A sparse proposal stage selects promising starts and ends per query and pairs them with no restriction on distance. A reranking head then scores each candidate using boundary evidence and span content.
Use cases and applications
The model is suited for legal and contract operations, healthcare and clinical documentation, financial services, insurance claims, customer support, and AI safety tooling. Specific applications include PII detection and redaction, contract clause extraction, knowledge graphs for agent memory, agent and model routing, guardrail classification, and clinical entity extraction with negation and dosage attributes.