File size: 10,218 Bytes
befca42
 
 
 
 
d74b5f7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
befca42
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d74b5f7
befca42
 
 
 
 
 
 
 
 
 
 
 
d74b5f7
 
 
 
 
befca42
 
 
 
 
 
 
d74b5f7
 
befca42
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
---
library_name: transformers
license: other
license_name: lfm1.0
license_link: LICENSE
language:
- ar
- zh
- en
- fr
- de
- hi
- id
- it
- ja
- ko
- pl
- pt
- ru
- es
- th
- vi
pipeline_tag: image-text-to-text
base_model: LiquidAI/LFM2.5-VL-3B
tags:
- liquid
- lfm2.5
- edge
- decision
- classification
- calibration
- system-one
- multimodal
---

<div align="center">
  <img
    src="https://cdn-uploads.huggingface.co/production/uploads/61b8e2ba285851687028d395/2b08LKpev0DNEk6DlnWkY.png"
    alt="Liquid AI"
    style="width: 100%; max-width: 100%; height: auto; display: inline-block; margin-bottom: 0.5em; margin-top: 0.5em;"
  />
  <div style="display: flex; justify-content: center; gap: 0.5em; margin-bottom: 1em;">
    <a href="https://playground.liquid.ai/"><strong>Try LFM</strong></a> •
    <a href="https://docs.liquid.ai/lfm/getting-started/welcome"><strong>Docs</strong></a> •
    <a href="https://discord.com/invite/liquid-ai"><strong>Discord</strong></a>
  </div>
</div>

# d1-3B

d1-3B is a 3B parameter **decision model** built on [LFM2.5-VL-3B](https://huggingface.co/LiquidAI/LFM2.5-VL-3B).
You give it a state (text, JSON, images, or a mix) and a set of named questions. It returns calibrated,
typed answers in **one forward pass with zero output tokens**: every answer is read directly from the
model's distribution over the options, with no generation and no parsing.

- **Best decision model under 10B**: 47.12 on the Decision Index 0.2.1. That is ahead of every 4B and 9B
  model on the public leaderboard and level with Decider 35B-A3B (47.11).
- **Multimodal**: images and text in the same state. It scores 73.7 on 11 public image benchmarks
  (LFM2.5-VL-3B: 73.9).
- **Fast**: 8 ms a decision on an NVIDIA RTX 4090, 9 ms on an AMD MI325X, 30 ms on an Apple M5 Pro.

![Decision Index against model size](assets/di_pareto.png)

> [!NOTE]
> 💻 **Demos**: Try d1-3B in a Hugging Face space without any setup:
> **TODO: add demo**


## 🗒️ Model Details

| Model | Parameters | Description |
|---|---|---|
| [LFM2.5-VL-3B](https://huggingface.co/LiquidAI/LFM2.5-VL-3B) | 3.1B | General-purpose vision-language model (base) |
| **[d1-3B](https://huggingface.co/LiquidAI/d1-3b-RC)** | 3.1B | Post-trained for single-pass, calibrated decisions |

d1-3B is a multimodal decision model with the following features:

- **Total parameters**: 3.12B
- **LM backbone**: LFM2.5-2.6B, 30 layers (22 double-gated short convolution blocks + 8 GQA)
- **Vision encoder**: SigLIP2 NaFlex shape-optimized 400M
- **Context length**: 32,768 tokens
- **Vocabulary size**: 128,000
- **Output**: no generated tokens. Each answer is a probability distribution over the question's options.

We recommend d1-3B wherever a pipeline needs a yes/no, a pick from named options, or a rating:
routing and triage, moderation, intent and topic classification, extraction checks, reranking, LLM-as-a-judge
scoring, agent guardrails, and visual inspection. It is not a chat model and does not write text.

## 🏃 How to use

Install the dependencies (requires `transformers>=5.14`):

```bash
pip install "transformers>=5.14" torch torchvision pillow
```

The model ships its own code, so load it with `trust_remote_code=True`:

```python
import io
import urllib.request

import torch
from PIL import Image
from transformers import AutoModel

device = "cuda" if torch.cuda.is_available() else "mps" if torch.backends.mps.is_available() else "cpu"
model = AutoModel.from_pretrained("LiquidAI/d1-3b-RC", trust_remote_code=True,
                                  dtype=torch.float32 if device == "cpu" else torch.bfloat16).to(device)

# Several named questions over one text state, answered in one pass
questions = {
    "refund": {"type": "noul", "instructions": "Is the customer asking for a refund?"},
    "team": {"type": "choice", "instructions": "Which team should handle this?",
             "criteria": {"billing": "Charges, refunds, invoices", "technical": "App or site faults",
                          "fraud": "Suspected unauthorised use"}},
    "urgency": {"type": "score", "instructions": "How urgent is this?",
                "criteria": ["Can wait", "Today", "Blocking the customer now"]},
}
print(model.system_one("I was charged twice this month, please refund one of them.", questions))

# An image as the whole state
url = "http://images.cocodataset.org/val2017/000000039769.jpg"  # two cats on a sofa
photo = Image.open(io.BytesIO(urllib.request.urlopen(url).read()))
print(model.system_one(None, {"cats": {"type": "choice", "instructions": "How many cats are there?",
                                       "criteria": {"one": "One", "two": "Two", "more": "Three or more"}}},
                       images=[photo]))

# Many requests, packed together with no padding
tickets = ["Where is my parcel? It was due Monday.", "The app crashes when I open settings."]
print(model.system_one_batch([(t, {"team": questions["team"]}) for t in tickets]))
```

| call | |
|---|---|
| `system_one(state, questions, images=None)` | Named questions over one state, in one pass. The state and its images are read once for all questions. |
| `system_one_batch([(state, questions[, images]), ...])` | Many requests, packed with no padding. |

A state is a string, any JSON value, or `None` when the images are the whole state.

### Questions and answers

Questions follow the Decision Index schema: `type`, `instructions`, and `criteria`.

| `type` | `criteria` | answer fields |
|---|---|---|
| `noul`: yes or no | optional: `{"true": "...", "false": "..."}` to define each side | `noul`: P(yes) |
| `choice`: one of named options | `{name: description}` | `choice`, `confidence`, `probabilities` |
| `score`: 2 to 10 ordered levels | a list of level descriptions, lowest first | `score` (the expected level), `confidence`, `probabilities`, `legend` |

Each call returns `{"answers": {name: answer}, "usage": {"input_tokens": n, "output_tokens": 0}}`.

## ⚡ Speed

Warm calls in bf16, median of 20 runs:

| | one question | 3 questions, one pass | 3.4k-token state | 384 px image | 64 states, packed |
|---|---:|---:|---:|---:|---:|
| NVIDIA RTX 4090 | **8.0 ms** | 21 ms | 102 ms | 17 ms | 475 / s |
| AMD MI325X | 9.2 ms | 14 ms | 44 ms | 18 ms | 1,106 / s |
| Apple M5 Pro (`mps`) | 30 ms | 41 ms | 0.64 s | 62 ms | 78 / s |

On NVIDIA GPUs, `model.compile(mode="reduce-overhead")` runs single questions as CUDA graphs. The RTX 4090
row uses it; without it a single question takes 16 ms. The first call with a new shape pays for kernel
selection or compilation, so warm up the shapes you serve.

## 📊 Performance

All results are on public benchmarks.

### Decision Index 0.2.1

The [Decision Index](https://huggingface.co/spaces/multimodalart/jev-decision-index) casts 38 public
benchmarks as 120,226 typed decisions, scored as chance-corrected skill. We scored d1-3B with the official
scorer; it is not a leaderboard submission. All other rows come from the public leaderboard
(snapshot 2026-09-28). Sizes are served parameters.

| Model | Size | Decision Index | Knowledge | Language | Retrieval | Tools | Arts |
|---|---:|---:|---:|---:|---:|---:|---:|
| Winnow-12B | 12B | 50.02 | 33.8 | 56.0 | 54.0 | 71.0 | 30.0 |
| **d1-3B** | **3B** | **47.12** | 27.8 | 53.5 | 52.2 | 66.7 | **34.6** |
| Decider 35B-A3B | 36B | 47.11 | 31.8 | 55.5 | 54.7 | 56.5 | 32.6 |
| JPT-9B | 9.7B | 46.89 | 31.7 | 56.7 | 44.6 | 67.0 | 28.6 |
| Decision 1.0 Lux | 9.7B | 43.49 | 30.9 | 48.0 | 50.0 | 57.2 | 26.4 |
| JPT-4B | 4.7B | 43.04 | 28.7 | 52.5 | 45.0 | 57.2 | 25.8 |
| Jet v6.2 | 4.7B | 42.60 | 28.7 | 43.9 | 48.2 | 62.9 | 27.0 |
| Decider 4B | 4.7B | 40.70 | 25.7 | 46.0 | 44.7 | 58.6 | 25.0 |
| Winnow-E4B | 8.0B | 39.89 | 22.3 | 45.1 | 43.8 | 62.5 | 22.8 |
| Decider 2B | 2.3B | 28.97 | 14.9 | 32.6 | 37.3 | 42.4 | 14.6 |

### Benchmarks as decisions

Evaluation splits are cast as typed decisions, with the same rows for every model. These scores follow the
decision protocol, not each benchmark's own.

| Benchmark | d1-3B | Decider 4B | Decider 2B |
|---|---:|---:|---:|
| SQuAD 2.0 | **83.3** | 76.0 | 67.7 |
| Civil Comments | 93.3 | 92.8 | **93.6** |
| MASSIVE intent | 86.9 | **88.3** | 81.1 |
| HelpSteer2 | 38.0 | **42.0** | 32.0 |
| PubMedQA | **68.3** | 63.3 | 65.7 |
| BoolQ | 86.3 | **89.0** | 87.3 |
| XNLI | 85.6 | **88.6** | 85.0 |
| PAWS-X | **76.4** | 69.8 | 59.5 |
| **Mean** | **77.3** | 76.2 | 71.5 |

d1-3B also scores 69.5 on [DecisionBench](https://huggingface.co/datasets/Hanno-Labs/decision-bench) (eng v1,
all 23,900 rows) and 68.8 on [Fast Decisions](https://huggingface.co/datasets/fastino/fast-decisions)
(dev split).

### Vision

Eleven public image benchmarks, read as decisions over each benchmark's options (at most 1,000 rows each),
compared with the base model:

| Benchmark | d1-3B | LFM2.5-VL-3B |
|---|---:|---:|
| AI2D | 80.1 | 80.9 |
| BLINK | 58.2 | 58.7 |
| CV-Bench | 82.7 | 87.6 |
| HallusionBench | 63.8 | 65.0 |
| MMBench | 84.3 | 84.3 |
| MME | 81.7 | 82.4 |
| MMStar | 59.4 | 61.2 |
| MMVP | 76.3 | 73.7 |
| POPE | 88.5 | 90.1 |
| VisualWebBench | 71.6 | 78.3 |
| VL-RewardBench | 63.6 | 50.9 |
| **Mean** | **73.7** | 73.9 |
| [ImajevBench](https://huggingface.co/datasets/mohit67890/imajev-bench) (dev and calibration, 253 rows) | 66.8 | 66.8 |

With the images removed, the same questions score 45.1, so the answers come from the images.

## ⚠️ Limitations

- **Knowledge-heavy questions**: larger models stay ahead on hard knowledge (for example MMLU-Pro and GPQA).
- **Thresholds**: calibrate any threshold on data from your own deployment before gating on it.
- **No text output**: d1-3B answers questions over fixed options. Use [LFM2.5-VL-3B](https://huggingface.co/LiquidAI/LFM2.5-VL-3B)
  for open-ended generation.

## 📬 Contact

- Got questions or want to connect? [Join our Discord community](https://discord.com/invite/liquid-ai)
- If you are interested in custom solutions with edge deployment, please contact [our sales team](https://www.liquid.ai/contact).

## Citation

```bibtex
@article{liquidai2025lfm2,
  title   = {LFM2 Technical Report},
  author  = {Liquid AI},
  journal = {arXiv preprint arXiv:2511.23404},
  year    = {2025}
}
```