Decision models · Fastino

GLiNER2 Base v1

Name: GLiNER2 Base v1
Author: Fastino

Task: Classifying and extracting
Returns: Labeled spans or fields
Precision tier: 68 / 100
Size: 210M
Max input: 2K tokens
License: apache-2.0
Released: 2025-07

What it decides

Returns: Entities, labels and structured fields from one schema, in one pass.
Checks: Entity extraction, classification and structured fields
Languages: English
Precision: 68 / 100 · Unifies extraction and classification; runs on a CPU (arXiv 2507.18546).
Clarity: Labeled spans or fields (13 of 15 clarity points)
How it runs: Labels spans or picks labels · reads the input once and writes nothing, so it answers in milliseconds.
License: apache-2.0 (Open (Apache, MIT))

Source: model card on Hugging Face (opens in a new tab) · How decision models are rated

How to run it

Original weights · Hugging Face

Full-precision files from Fastino, for Transformers, Sentence Transformers, Text Embeddings Inference, Infinity or vLLM.

huggingface.co/fastino/gliner2-base-v1 (opens in a new tab)

No sign-in needed.

Links checked 2 days ago against Hugging Face, Ollama and GitHub.

Apps that run GLiNER2 Base v1

Our pick for each kind of computer, assuming the model fits in its memory. Pick your own GPU or Mac in the Model Finder for an exact answer.

Mac (Apple silicon)

GLiNER

Made for this model family: every label is scored in one pass, fast even on a CPU.

Windows PC, NVIDIA GPU

GLiNER

Made for this model family: every label is scored in one pass, fast even on a CPU.

Windows PC, AMD Radeon

GLiNER

Can only use the CPU on this hardware, so it will be slow.

Linux, NVIDIA GPU

GLiNER

Made for this model family: every label is scored in one pass, fast even on a CPU.

Linux server, 80 GB+ GPUs

GLiNER

Made for this model family: every label is scored in one pass, fast even on a CPU.

App	Type	Opens it as	Systems	Engine	Start
GLiNER	Library	Original weights	macOSWindowsLinux	PyTorch, ONNX	`pip install gliner2`

Which apps can open a model depends on the files published for it: GGUF for llama.cpp-based apps, MLX for Apple’s engine, the original weights for vLLM and SGLang. Compare every app.

Memory and speed

Each precision on an NVIDIA GeForce RTX 4090, for one decision on a 256-token input. The best hardware table below covers other GPUs and Macs.

Format	Memory needed	Time per decision	Decisions per second	Quality	Download
`Q8_0`	0.7 GB	5.0 ms	201/s	68Good quality (B) · 68 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	All files (opens in a new tab)
`F16`	0.9 GB	5.1 ms	195/s	68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	Original weights (opens in a new tab)

One request at a time. Servers that batch (Text Embeddings Inference, vLLM, Infinity) handle many more decisions per second.

Best hardware for GLiNER2 Base v1

The least expensive GPU in our catalog that runs it without offloading is the Intel Arc A380 (launch price $139) at F16.

Quality	Hardware	Quant	Memory	Time per decision	Runs
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	AMD Radeon RX 7900 XTX	`F16`	1.1 GB / 24.0 GB	4.9 ms	Runs great
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	AMD Radeon RX 9070 XT	`F16`	1.1 GB / 16.0 GB	5.1 ms	Runs great
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	AMD Ryzen AI Max+ 395 (128 GB)	`F16`	1.1 GB / 96.0 GB	5.8 ms	Runs great
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	Apple M1 (16 GB)	`F16`	1.1 GB / 10.7 GB	46 ms	Runs great
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	Apple M1 Max (64 GB)	`F16`	1.1 GB / 48.0 GB	14 ms	Runs great
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	Apple M2 Ultra (192 GB)	`F16`	1.1 GB / 144 GB	8.0 ms	Runs great
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	Apple M3 Ultra (256 GB)	`F16`	1.1 GB / 192 GB	7.8 ms	Runs great
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	Apple M3 Ultra (512 GB)	`F16`	1.1 GB / 384 GB	7.8 ms	Runs great
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	Apple M4 (16 GB)	`F16`	1.1 GB / 10.7 GB	29 ms	Runs great
68Good quality (B) · 68 of 100: how capable the model is at F16, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking.	Apple M4 (24 GB)	`F16`	1.1 GB / 16.1 GB	29 ms	Runs great

Check it on your own hardware

Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.