Model cardFastino · 2025-07
GLiNER2 paper (arXiv 2507.18546)
Measured on: Seven zero-shot classification sets
- Zero-shot classification average
- 0.72 (Banking77 0.70, SST-2 0.86, AG News 0.74)
Peer-visible paper; the 205M model.
GLiNER2 base
FastinoOpen weights
The smaller v1-generation GLiNER2 checkpoint.
- typed-decisions
- —
- Metatext suite
- 0.631
- ECE untuned
- 0.350
- p50 per case
- 190 ms
not run
3/7 over baseline
four CPU cores
Specifications
- Checkpoint
- fastino/gliner2-base-v1
- Publisher
- Fastino
- Mechanism
- Schema-conditioned encoder (extraction + classification)
- Backbone
- DeBERTa-v3 (span grid)
- Parameters
- 205M
- License
- Apache 2.0
- Runs as
- open weights, self-hosted
- Library
- gliner2
- Primitives
- Choice, Score, Noul, Extraction
- Pricing
- none — your compute
Metatext decisionsrank 5 of 21
Full leaderboard →Per-question accuracy — default payload, parallel mode, 120 judgments
| Task | Question | Type | Accuracy | Baseline | Beats? |
|---|---|---|---|---|---|
| agent_traces | outcome | choice | 0.571 | 0.357 | ✅ |
| agent_traces | needs_human | noul | 0.714 | 0.714 | ❌ |
| customer_service | department | choice | 0.850 | 0.400 | ✅ |
| customer_service | refund_requested | noul | 0.900 | 0.700 | ✅ |
| customer_service | frustration | score | 0.500 | 0.500 | ❌ |
| security_incidents | severity | choice | 0.312 | 0.312 | ❌ |
| security_incidents | actively_exploited | noul | 0.688 | 0.688 | ❌ |
Published numbers
Reported by the authors or by third parties, on their own suites. Not comparable to the leaderboard above; attributed so you can check them.
Frequently asked questions
How does GLiNER2 base compare to Jev?
On the Metatext suite GLiNER2 base scores 0.631 against Jev's 0.863.
Is GLiNER2 base open source?
Yes. fastino/gliner2-base-v1 is released under Apache 2.0 and runs on your own hardware with the gliner2 library.
Can GLiNER2 base extract entities?
Yes. GLiNER2 base has a span head and scores 0.935 micro-F1 on the eight-case extraction task.