Jev directory
CommunityJev Models & Reimplementations
GLiNER2.5-Decide
Fastino's 340M English classification model and the headline member of the GLiNER2.5 family: pass any label set at call time and get intent, routing, sentiment or multi-label tags back in one forward pass, at 60.2% exact-match average on the Fast Decisions suite.
- Reading time
- 2 min
- Published by
- Community
- Author
- Fastino
- Last verified
- Oct 2026
- Languages
- Python
- Capabilities
- Open sourceSelf-hostedRuns locally
Highlights
- 340M-parameter English classifier; the card's headline model and the top row of its Fast Decisions table at 60.2%.
- Labels are supplied at call time and scored in a single forward pass, with no prompt template and no generated tokens.
- One call can score several heads at once; single-label tasks return one string, multi-label every label above threshold.
- Load with AutoExtractor after pip install "gliner2[local]"; the card says to ship it locally.
- Built on fastino/gliner2-large-v1 under Apache-2.0; 115,972 downloads and 417 likes at review time.
Quickstart
python
from gliner2 import AutoExtractor
model = AutoExtractor.from_pretrained("fastino/GLiNER2.5-Decide")Watch out
Card reviewed 2026-10-17; the 60.2% exact-match average is self-reported on the card and the comparison suite is English. The card states the model does not reason, explain or answer open questions, and no independent rerun of its benchmark table is cited.
Questions about this entry
Short answers drawn from the linked project or page.
The 340M English classification model in Fastino's GLiNER2.5 family. Text plus any label set goes in; a label or label set comes out of a single forward pass, with no prompt template, no generated tokens and no fixed taxonomy.
It follows the same principle as Jev's System One: decisions are scored against supplied options rather than generated as text. The card documents a local AutoExtractor Python interface rather than a /v1/systemone endpoint, so it is a family relative, not a drop-in Jev server.
Intent, routing, sentiment, priority, policy and multi-label tags, with documented examples for customer support, banking, travel and clinic requests, plus a passage question, book classification, labels with descriptions and an ordinal scale.
On fastino/fast-decisions — 17 domains, 300 held-out examples each — it averages 60.2% exact-match, ahead of GLiNER2.5-Decide-1B 59.6%, JevK5 57.6%, GLiNER2.5-multi-Decide 56.7%, SemIf 56.4%, GLiFormer 49.0% and Laya Router 46.6%. The numbers are the card's own.
The card says to use this 340M model for English-only text and GLiNER2.5-multi-Decide when the input is multilingual; GLiNER2.5-Decide-1B is the larger English option at 59.6%, slightly below this model's 60.2%.
pip install "gliner2[local]", then AutoExtractor.from_pretrained("fastino/GLiNER2.5-Decide") and model.classify_text(text, {"intent": [...]}). Single-label heads return one string; multi-label heads return every label above the threshold.
Apache-2.0. The base model is fastino/gliner2-large-v1, and the card belongs to the GLiNER2.5 line with the arXiv paper 2507.18546 and the fastino-ai/GLiNER2 GitHub repository linked.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: Standard One SH (StandardThinking)
LinkedIn: What multimodal decision models actually change, and what they do not.
LinkedInJevDirectory.org
Hacker News: Decision Index 0.2.1 scores open reproductions of Jev on a frozen suite.
jackbravo4 pointsDecision Index 0.2.1 – 70 open reproductions of TypeSafe Jev's Decision ModelRead the thread on Hacker NewsHacker News: Pointer to a DiffusionGemma plus vLLM tutorial for a Jev-shaped classifier.
1 pointHow to build a Jev-style classifier with DiffusionGemma and vLLMRead the thread on Hacker NewsMore Jev resources like this
Fastino's 1B English member of the GLiNER2.5 classification family: pass any label set at call time for intent, routing, sentiment or multi-label tags and get labels back in a single forward pass with no generated tokens. It averages 59.6% exact-match on Fast Decisions.
Jev Models & Reimplementations#community#open-models#local
Community
The 322M mmBERT-base build of Laya for 100+ languages: state plus typed questions return typed answers with probabilities in one forward pass, with up to 8,192 tokens for long documents.
Jev Models & Reimplementations#community#open-models#local
Community
The Hugging Face hub for Laya's open System 1 checkpoints: the 421M English model at the repo root, a 322M multilingual build in a subfolder, and the typed-decisions fine-tune, all Apache-2.0.
Jev Models & Reimplementations#community#open-models#local
Community
A DistilBERT prototype from akumar0205 that sorts historical email into legitimate and spam, reporting 0.987 accuracy and 0.983 spam F1 on a held-out SpamAssassin split; it is not a phishing or malware detector.
Jev Models & Reimplementations#community#open-models#local
Community
Fastino's 287M multilingual classification model in the GLiNER2.5 family: pass any label set at call time and get labels back in a single forward pass, with no prompt template and no generated tokens. It averages 56.7% exact-match on the Fast Decisions suite.
Jev Models & Reimplementations#community#open-models#local
Community
TokenRhythm's 4B prefill-only decision model built on NeoHorse-1-4B: given a state and application-defined questions it returns Choice, Noul, and Score probabilities without generating text, reporting a 77.70 six-group aggregate under Apache-2.0.
Jev Models & Reimplementations#community#open-models#local
Community
From the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.