File size: 8,213 Bytes
7c62260
55a0fff
 
 
 
 
 
 
 
 
 
 
 
7c62260
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a98e19b
 
 
 
 
 
 
 
 
 
 
7c62260
 
 
a98e19b
 
 
 
 
 
7c62260
4106286
55a0fff
4106286
 
 
 
 
 
 
 
 
 
 
 
1d533f8
 
69c7036
1d533f8
 
 
4106286
 
 
 
 
 
 
 
 
 
 
55a0fff
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4106286
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1d533f8
 
4106286
1d533f8
 
 
 
 
 
69c7036
 
 
1d533f8
 
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
---
license: cc-by-4.0
task_categories:
- text-generation
language:
- en
tags:
- chemistry
- supramolecular
- host-guest
- molecular-recognition
- llm-benchmark
pretty_name: 'SupraBench: Binding Affinity Prediction (BAP)'
dataset_info:
  features:
  - name: id
    dtype: string
  - name: task
    dtype: string
  - name: answer
    dtype: float64
  - name: host_name
    dtype: string
  - name: molecule
    dtype: string
  - name: version
    dtype: string
  - name: question
    dtype: string
  - name: source
    dtype: string
  - name: prompt_strategy
    dtype: string
  splits:
  - name: base
    num_bytes: 4589535
    num_examples: 2609
  - name: cot
    num_bytes: 4589535
    num_examples: 2609
  - name: fewshot
    num_bytes: 4589535
    num_examples: 2609
  download_size: 1533644
  dataset_size: 13768605
configs:
- config_name: default
  data_files:
  - split: base
    path: data/base-*
  - split: cot
    path: data/cot-*
  - split: fewshot
    path: data/fewshot-*
---


# SupraBench

**SupraBench** is the first benchmark for evaluating large language models on
**supramolecular host-guest chemistry** reasoning. It comprises four fundamental
tasks plus an auxiliary vision task, and ships a domain text corpus for
domain-adaptive pretraining (DAPT).

> Supramolecular chemistry studies non-covalent host-guest assemblies that
> underpin drug delivery, chemical sensing, and in-vivo toxin sequestration.
> Designing host-guest systems is slow (days of dry-lab verification per pair);
> SupraBench probes whether LLMs can reason about these systems directly.

## Links

- 📄 **Paper:** [`arXiv:2606.13477`](https://arxiv.org/abs/2606.13477)
- 💻 **Code:** https://github.com/Tianyi-Billy-Ma/SupraBench
- 🤗 **All datasets:** https://huggingface.co/SupraBench

## The dataset family

| Dataset | Task | Description |
|---|---|---|
| [`SupraBench/bap`](https://huggingface.co/datasets/SupraBench/bap) | Binding Affinity Prediction | regress log K_a for a host-guest pair |
| [`SupraBench/tbs`](https://huggingface.co/datasets/SupraBench/tbs) | Top-Binder Selection | pick the strongest binder among 4 candidate guests |
| [`SupraBench/sid`](https://huggingface.co/datasets/SupraBench/sid) | Solvent Identification | 6-way solvent classification from structure |
| [`SupraBench/hgd`](https://huggingface.co/datasets/SupraBench/hgd) | Host-Guest Description | open-ended QA on host/guest property profiles |
| [`SupraBench/EU-PMC`](https://huggingface.co/datasets/SupraBench/EU-PMC) | Text corpus | 16M-token supramolecular corpus for DAPT |
| [`SupraBench/Binding-Affinity`](https://huggingface.co/datasets/SupraBench/Binding-Affinity) | Comprehensive anchor | per-record binding data + host/guest SMILES, 2D, 3D, environment |

Each task dataset is partitioned by prompting strategy into three splits:
`base`, `cot`, and `fewshot` (the strategy is also recorded in the
`prompt_strategy` column). The CB[7] generalization records are a subset of every
split and remain identifiable by host name.

## Usage

```python
from datasets import load_dataset

# prompting-strategy split: "base" | "cot" | "fewshot"
ds = load_dataset("SupraBench/bap", split="cot")
print(ds[0]["question"])   # fully rendered prompt
print(ds[0]["answer"])     # reference label
```

To run the benchmark end-to-end (inference + scoring) see the
[code repository](https://github.com/Tianyi-Billy-Ma/SupraBench):

```bash
uv run python src/main.py \
    --task-config  configs/tasks/bap_base.yaml \
    --model-config configs/models/openrouter_qwen35_27b.yaml \
    --output-dir   outputs/
```

## Dataset statistics

| Task | # Samples |
|---|---|
| BAP | 2,609 |
| TBS | 2,264 |
| SID | 2,172 |
| HGD | 135 |

Top-4 hosts (BAP / TBS / SID counts): CB[8] 261/200/571, CB[7] 217/200/217,
beta-CD 201/200/264, p-SC4 144/144/225.

`SupraCorpus` (EU-PMC): 420,950 raw filtered articles -> 133,867 high-precision
articles -> ~16M tokens.

## Performance report

Main results from the SupraBench paper across the four fundamental tasks
(8 LLMs x 3 prompting strategies). **Bold** = best, *italic* = second-best per
column. Arrows give the optimization direction.

### Base

| Model | BAP MAE down | BAP RMSE down | TBS ACC up | TBS Regret down | SID F1 up | SID B.Acc up | HGD Recall up | HGD Prec up | HGD F1 up |
|---|---|---|---|---|---|---|---|---|---|
| Qwen3.5-9B | 2.491 | 3.360 | 0.379 | 0.930 | 0.159 | 0.166 | 0.040 | 0.023 | 0.043 |
| Qwen3.5-27B | 1.803 | 2.503 | 0.404 | 0.851 | 0.225 | 0.364 | 0.495 | 0.072 | 0.122 |
| Llama3.1-8B | 2.699 | 3.630 | 0.228 | 1.281 | 0.151 | 0.225 | 0.266 | 0.059 | 0.092 |
| Llama3.1-70B | 1.632 | 2.149 | 0.338 | 1.054 | 0.118 | 0.254 | 0.487 | **0.091** | **0.152** |
| GPT-5.4-Mini | 1.549 | 2.182 | 0.428 | 0.810 | 0.219 | 0.274 | 0.437 | 0.086 | 0.137 |
| GPT-5.4-Nano | 1.642 | 2.169 | 0.411 | 0.816 | 0.182 | 0.347 | 0.472 | 0.062 | 0.107 |
| Gemini-3-Flash | **1.248** | **1.679** | **0.498** | **0.647** | **0.350** | **0.470** | **0.506** | 0.067 | 0.118 |
| DeepSeek-v4 | *1.433* | *1.994* | *0.461* | *0.730* | *0.309* | *0.381* | *0.500* | *0.090* | *0.141* |

### Few-Shot

| Model | BAP MAE down | BAP RMSE down | TBS ACC up | TBS Regret down | SID F1 up | SID B.Acc up | HGD Recall up | HGD Prec up | HGD F1 up |
|---|---|---|---|---|---|---|---|---|---|
| Qwen3.5-9B | 3.650 | 4.820 | 0.370 | 0.951 | 0.154 | 0.150 | 0.000 | 0.022 | 0.042 |
| Qwen3.5-27B | 2.258 | 3.256 | 0.392 | 0.889 | 0.178 | 0.257 | 0.636 | **0.585** | **0.580** |
| Llama3.1-8B | 5.504 | 6.940 | 0.283 | 1.227 | 0.142 | 0.182 | 0.655 | 0.369 | 0.456 |
| Llama3.1-70B | 1.774 | 2.359 | 0.354 | 1.026 | 0.144 | 0.185 | 0.631 | *0.474* | *0.531* |
| GPT-5.4-Mini | 1.958 | 2.808 | 0.430 | 0.824 | 0.141 | *0.291* | 0.542 | 0.228 | 0.307 |
| GPT-5.4-Nano | 2.176 | 2.894 | 0.419 | 0.819 | 0.190 | 0.270 | 0.532 | 0.095 | 0.152 |
| Gemini-3-Flash | **1.257** | **1.702** | **0.513** | **0.619** | **0.389** | **0.421** | *0.660* | 0.364 | 0.448 |
| DeepSeek-v4 | *1.618* | *2.276* | *0.470* | *0.713* | *0.203* | 0.225 | **0.720** | 0.303 | 0.352 |

### CoT

| Model | BAP MAE down | BAP RMSE down | TBS ACC up | TBS Regret down | SID F1 up | SID B.Acc up | HGD Recall up | HGD Prec up | HGD F1 up |
|---|---|---|---|---|---|---|---|---|---|
| Qwen3.5-9B | 3.664 | 4.885 | 0.382 | 0.944 | 0.167 | 0.197 | 0.300 | 0.039 | 0.068 |
| Qwen3.5-27B | 2.438 | 3.468 | 0.398 | 0.898 | 0.254 | *0.415* | **0.526** | 0.051 | 0.092 |
| Llama3.1-8B | 4.911 | 6.279 | 0.293 | 1.220 | 0.154 | 0.153 | 0.380 | **0.102** | **0.144** |
| Llama3.1-70B | 1.833 | 2.512 | 0.373 | 0.985 | 0.106 | 0.380 | 0.421 | 0.055 | 0.097 |
| GPT-5.4-Mini | 2.036 | 2.887 | 0.429 | 0.828 | 0.220 | 0.282 | 0.444 | 0.080 | 0.129 |
| GPT-5.4-Nano | 2.160 | 2.881 | 0.410 | 0.822 | 0.174 | 0.257 | 0.492 | 0.056 | 0.098 |
| Gemini-3-Flash | **1.261** | **1.723** | **0.510** | **0.609** | **0.331** | **0.432** | 0.512 | 0.062 | 0.110 |
| DeepSeek-v4 | *1.541* | *2.183* | *0.445* | *0.743* | *0.307* | 0.414 | *0.522* | *0.080* | *0.134* |

Takeaways: frontier proprietary LLMs (Gemini-3-Flash, DeepSeek-v4) lead the
quantitative tasks, yet every task leaves substantial headroom; no single
prompting strategy is universally best; and CoT amplifies rather than fixes the
underlying reasoning gap on binding-affinity prediction.

## Sources & license

Binding records are derived from [SupraBank](https://suprabank.org/) (CC-BY-4.0);
the text corpus is built from open-access [Europe PMC](https://europepmc.org/)
articles subject to each article's individual license; molecular structures use
[PubChem](https://pubchem.ncbi.nlm.nih.gov/) and
[OPSIN](https://github.com/dan2097/opsin).

## Citation

If you use SupraBench, please cite the paper and the upstream data sources.

```bibtex
@article{ma2026suprabench,
  title   = {SupraBench: A Benchmark for Supramolecular Host--Guest Chemistry Reasoning in Large Language Models},
  author  = {Ma, Tianyi and Ma, Yijun and Wang, Zehong and Sun, Weixiang and Li, Ziming and Schmidt, Connor R. and Zhang, Chuxu and Webber, Matthew J. and Ye, Yanfang},
  year    = {2026},
  eprint        = {2606.13477},
  archivePrefix = {arXiv},
  journal = {arXiv preprint arXiv:2606.13477}
}
```