Open source, self-hosted. There is no hosted API. You download the weights and run them on your own hardware. How open-source models are listed.
What GLiNER2.5-Decide is
GLiNER2.5-Decide is Fastino Labs’ System One model: an encoder-based decision model rather than a generative one. It evaluates a set of user-defined typed questions and rules against a piece of text in one pass, then jointly decodes the answers, returning probability distributions and confidence scores. Beyond typed Choice, Score and Noul questions, Fastino describes it returning character-level spans, relations, and structured records, and applying implications, exclusions, cardinality limits, and ordinal bounds across related decisions in one call.
The model card is direct about scope: “This release is not a general-purpose model. It does not reason, explain, or answer open questions.”
How it is built
The Hugging Face model card lists GLiNER2.5-Decide as a 340M-parameter model fine-tuned from Fastino’s own gliner2-large-v1, a DeBERTa-v3-large encoder. The Hugging Face API’s safetensors metadata for the published weights reports 486,444,053 parameters, which does not match the 340M figure. Neither source explains the gap.
It loads through the gliner2 Python package (AutoExtractor.from_pretrained("fastino/GLiNER2.5-Decide")) and accepts label sets at runtime, so a new task needs no retraining or prompt template.
What it’s good at, per Fastino’s own benchmark
Fastino evaluated the model on “Fast Decisions,” an internal suite of 17 datasets covering intent routing, triage, sentiment and content understanding. This is Fastino’s own benchmark, not an independently run one. The blog post and company X post give GLiNER2.5-Decide 60.1% average accuracy, ahead of JevK5 (57.5%), SemIf (56.4%) and Laya (46.6%), leading on 9 of 17 datasets. The Hugging Face card’s own results table gives slightly different figures for the same run: 60.2% and JevK5 at 57.6%. Fastino highlights support intent classification at 75.3%, 18.6 points ahead of the next model, and banking intent at 64.3%, 8.6 points ahead.
The GitHub repository behind it, fastino-ai/GLiNER2, had 2,169 stars and was last pushed on 24 September 2026, the day of release.
What an independent benchmark shows
The Jev Decision Index by multimodalart is a broader test than Fastino’s. Its edition 0.4.1, in the file generated on 12 November 2026, ranks entries by a Full score built only from private tests, 60% on the same skills and 40% in new domains, on a scale where 0 is random guessing and 100 is perfect. GLiNER2.5-Decide scores 10.14 against Jev 1.14’s 60.24, and 11.48 against Jev 1.14’s 60.7 on the public benchmarks alone. The headline changed in 0.3 and again in 0.4, so these numbers do not compare with the 0.2.1 figures this page gave before, or with its 0.3.1 Full score of 9.24. It is not the top model under 500M parameters on the board: smaller entries such as Fastino’s own GLiDE 0.4B at 33.84, GLiClass v4.1 edge (0.4B) at 32.47 and Bekko System One v0 400M at 17.98 score higher on the Full score. Fastino’s suite ranks JevK5 below GLiNER2.5-Decide, but the Index’s JevK5 v0.3 entry scores 36.59. Its expected calibration error, the average gap between stated confidence and actual accuracy, is 0.088 against Jev 1.14’s 0.080.
Access and running it
Weights are Apache 2.0 on Hugging Face with no waitlist. Fastino says the model is “efficient enough to run locally on consumer-grade CPUs or in air-gapped environments,” and co-founder George Maloney gives CPU latency around 167ms and GPU latency of 38 to 47ms for short requests. Fastino also runs hosted inference and fine-tuning at http://agent.fastino.ai, for calling from inside a coding agent; no pricing for that service appeared in the sources checked.
What it’s not for
It does not generate text, reason through multi-step problems, or answer open-ended questions; the model card says so outright. No context-window limit, maximum Choice count, or per-call question cap is documented anywhere checked, so those limits are unknown rather than unlimited.
Specifications
| System One Score | 27.2 from 2 of 3 boardsJev Decision Index 13.8, JevBench 40.6. How the score works |
| Inputs | Text |
| Question types | ChoiceScoreNoul |
| Max Choice options | Not documented |
| Score levels | Not documented |
| Questions per call | Not documented |
| Total context | Not documented |
| State budget | Not documented |
| Rate limit | None for self-hosted use. Fastino's hosted inference and fine-tuning API at http://agent.fastino.ai did not have published rate limits in the sources checked. |
| Endpoint | http://agent.fastino.ai |
| Licence | Apache-2.0 |
| SDKs | Python: gliner2 |
Neither the blog post, the Hugging Face model card, nor the GitHub repo README documented a fixed context window, a maximum number of Choice options, or a cap on questions per call as of this check.
Versions
- fastino/GLiNER2.5-Decide, 24 Sep 2026, 340M-parameter encoder fine-tuned from fastino/gliner2-large-v1, a DeBERTa-v3-large base per the Hugging Face model card. The Hugging Face API's safetensors metadata reports 486,444,053 total parameters for the published weights, which does not match the 340M figure stated in the blog post and model card; this discrepancy is unresolved as of this check. Apache 2.0 license. Release notes
Use cases
What people use GLiNER2.5-Decide for, one page per pattern.
Support inbox triage with System One models
Send a support ticket to Jev once with every question attached. Category comes back as a selected label, severity and frustration as numbers on scales you wrote, refund intent as a probability. Your code reads those values and decides what happens to the ticket.
Intent and model routing with System One models
One Jev call reads an incoming request and returns its intent as a label plus a difficulty rating on a scale you wrote. Your router reads both numbers and picks the handler: deterministic code, a cheap model, an expensive one, or a human queue.
Agent routing and skill selection with System One models
An agent choosing from a long skill roster reads one truncated line per entry and often loads the wrong thing. Jev ranks every entry in one request and separately answers whether any skill applies at all, so the agent gets a short hint instead of a guess.
Structured extraction with System One models
There are two routes. Clairic, Cloudflare's extraction model, takes named fields and returns typed values, each with the probability that it is missing. Jev cannot write values, so code finds candidate spans, a choice question picks one, and code copies it unchanged.
Examples built with GLiNER2.5-Decide
The most-starred and most-viewed entries in the directory. All GLiNER2.5-Decide examples.

GLiNER2.5-Decide
Open-weights 340M-parameter decision model from Fastino Labs. It jointly decodes typed classification questions and, per Fastino's own internal benchmark, leads 9 of 17 datasets against JevK5, SemIf and Laya.
2.2k starsCPU latency 167 msGPU latency 38-47 ms

Jev Decision Index: Jev against 50+ open decision models
A leaderboard that runs Jev and 54 open System One models and clones through the same 120,000-question suite on one RTX PRO 6000, scoring accuracy and calibration. Jev leads the 0.2 edition at 51.67 on a chance-corrected scale, with AutoJev-27B close behind at 50.94.
Benchmarks
Published scores, each linked to its source. A row marked vendor-run was measured by the vendor itself, so read it as one side's result. The compare hub sets these beside the other models.
| Benchmark | Metric | Score | Run by | Date |
|---|---|---|---|---|
| Jev Decision Index 0.4.1 | Full score | 10.14 score | Independent | 12 Nov 2026 |
| Jev Decision Index 0.4.1 | Public score | 11.48 score | Independent | 12 Nov 2026 |
| Jev Decision Index 0.4.1 | Expected calibration error | 0.0884 ECE | Independent | 12 Nov 2026 |
| Jev Decision Index 0.3.1 | Full score | 9.24 score | Independent | 10 Oct 2026 |
| Jev Decision Index 0.3.1 | Public score | 11.48 score | Independent | 10 Oct 2026 |
| Jev Decision Index 0.3.1 | Expected calibration error | 0.0884 ECE | Independent | 10 Oct 2026 |
| JevBench 1.8.10 | Capability Score | 34.6 score | Independent | 5 Oct 2026 |
| Fast Decisions | Average accuracy | 60.1 % | vendor-run | 24 Sep 2026 |
- Jev Decision Index, Full score: Row: GLiNER2.5-Decide. Counts 60% private same-skill tests and 40% private new-domain tests; public benchmarks no longer count. 0 is random guessing, 100 is perfect.
- Jev Decision Index, Public score: Row: GLiNER2.5-Decide.
- Jev Decision Index, Expected calibration error: Row: GLiNER2.5-Decide. Lower is better.
- Jev Decision Index, Full score: Row: GLiNER2.5-Decide. Counts 20% public benchmarks and 80% private tests; 0 is random guessing, 100 is perfect.
- Jev Decision Index, Public score: Row: GLiNER2.5-Decide.
- Jev Decision Index, Expected calibration error: Row: GLiNER2.5-Decide. Lower is better.
- JevBench, Capability Score: Row: Fastino GLiNER-2.5-Decide (hosted API). Intelligence 7.0, Calibration 62.1. 37th by Capability Score on the hosted API board. Board release v1.8.10, read 2026-11-15.
- Fast Decisions, Average accuracy: Fastino's own suite of 17 datasets and 5,100 test examples; the Hugging Face card gives 60.2.
Comparisons
Head-to-head pages that put GLiNER2.5-Decide against other models. Every model sits in one table on the compare hub.
Jev vs GLiNER: typed decisions against span extraction
Classic GLiNER extracts entity spans. Fastino's GLiNER2.5-Decide, released 2026-09-24, answers typed decisions instead. Both differ from Jev in weights and benchmarks.
Updated