# Formal Language Generation on SMILES-Bench

Source: [Formal Language Generation on SMILES-Bench](<https://www.sota2.com/research/sota/formal-language-generation-on-smiles-bench>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T06:05:48.165Z

Evaluation of model performance on generating formally correct SMILES strings using the SMILES-Bench dataset, measured by syntactic accuracy at various top-k cutoffs (k=1, 3, 5, 10).

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Formal Language Generation |
| Dataset (including any stated split) | SMILES-Bench |
| Metric | Syntactic Accuracy @1 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 16 of 16 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Syntactic Accuracy @1](<https://www.sota2.com/research/sota/formal-language-generation-on-smiles-bench.md?metric=Syntactic%20Accuracy%20%401>): Higher is better; unit: PERCENT.

- [Syntactic Accuracy @3](<https://www.sota2.com/research/sota/formal-language-generation-on-smiles-bench.md?metric=Syntactic%20Accuracy%20%403>): Higher is better; unit: PERCENT.

- [Syntactic Accuracy @5](<https://www.sota2.com/research/sota/formal-language-generation-on-smiles-bench.md?metric=Syntactic%20Accuracy%20%405>): Higher is better; unit: PERCENT.

- [Syntactic Accuracy @10](<https://www.sota2.com/research/sota/formal-language-generation-on-smiles-bench.md?metric=Syntactic%20Accuracy%20%4010>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Syntactic Accuracy @1 | Result date | Report paper |
| --- | --- | --- | --- |
| LAVE | 99.4 | Not recorded | 69818109a8b01e76cc8e225e |
| LAVE | 98.8 | Not recorded | 69818109a8b01e76cc8e225e |
| LAVE | 98.7 | Not recorded | 69818109a8b01e76cc8e225e |
| LAVE | 96.6 | Not recorded | 69818109a8b01e76cc8e225e |
| IG-CD | 96.4 | Not recorded | 69818109a8b01e76cc8e225e |
| FS-CD | 96 | Not recorded | 69818109a8b01e76cc8e225e |
| IG-CD | 95.1 | Not recorded | 69818109a8b01e76cc8e225e |
| IG-CD | 94.3 | Not recorded | 69818109a8b01e76cc8e225e |
| IG-CD | 93.7 | Not recorded | 69818109a8b01e76cc8e225e |
| FS-CD | 93.1 | Not recorded | 69818109a8b01e76cc8e225e |
| FS-CD | 78.3 | Not recorded | 69818109a8b01e76cc8e225e |
| FS-CD | 75.7 | Not recorded | 69818109a8b01e76cc8e225e |
| NO-CD | 74 | Not recorded | 69818109a8b01e76cc8e225e |
| NO-CD | 70.1 | Not recorded | 69818109a8b01e76cc8e225e |
| NO-CD | 66 | Not recorded | 69818109a8b01e76cc8e225e |
| NO-CD | 62.3 | Not recorded | 69818109a8b01e76cc8e225e |

### LAVE

Syntactic Accuracy @1: 99.4. Result ID: 69965a925698665ec4a01e6a.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

Recorded evaluation configuration: {"Model backbone":"Dream-7B"}. This record does not assert independent reproduction.

### LAVE

Syntactic Accuracy @1: 98.8. Result ID: 69965a925698665ec4a01e6e.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

Recorded evaluation configuration: {"Model backbone":"DiffuC-7B"}. This record does not assert independent reproduction.

### LAVE

Syntactic Accuracy @1: 98.7. Result ID: 69965a925698665ec4a01e62.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

Recorded evaluation configuration: {"Model backbone":"LLaDA-8B"}. This record does not assert independent reproduction.

### LAVE

Syntactic Accuracy @1: 96.6. Result ID: 69965a925698665ec4a01e66.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

Recorded evaluation configuration: {"Model backbone":"LLaDA-1.5"}. This record does not assert independent reproduction.

### IG-CD

Syntactic Accuracy @1: 96.4. Result ID: 69965a925698665ec4a01e6d.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ab01e209131cea849f1.md>)

Recorded evaluation configuration: {"Model backbone":"DiffuC-7B"}. This record does not assert independent reproduction.

### FS-CD

Syntactic Accuracy @1: 96. Result ID: 69965a925698665ec4a01e6c.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

Recorded evaluation configuration: {"Model backbone":"DiffuC-7B"}. This record does not assert independent reproduction.

### IG-CD

Syntactic Accuracy @1: 95.1. Result ID: 69965a925698665ec4a01e69.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ab01e209131cea849f1.md>)

Recorded evaluation configuration: {"Model backbone":"Dream-7B"}. This record does not assert independent reproduction.

### IG-CD

Syntactic Accuracy @1: 94.3. Result ID: 69965a925698665ec4a01e61.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ab01e209131cea849f1.md>)

Recorded evaluation configuration: {"Model backbone":"LLaDA-8B"}. This record does not assert independent reproduction.

### IG-CD

Syntactic Accuracy @1: 93.7. Result ID: 69965a925698665ec4a01e65.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ab01e209131cea849f1.md>)

Recorded evaluation configuration: {"Model backbone":"LLaDA-1.5"}. This record does not assert independent reproduction.

### FS-CD

Syntactic Accuracy @1: 93.1. Result ID: 69965a925698665ec4a01e68.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

Recorded evaluation configuration: {"Model backbone":"Dream-7B"}. This record does not assert independent reproduction.

### FS-CD

Syntactic Accuracy @1: 78.3. Result ID: 69965a925698665ec4a01e64.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

Recorded evaluation configuration: {"Model backbone":"LLaDA-1.5"}. This record does not assert independent reproduction.

### FS-CD

Syntactic Accuracy @1: 75.7. Result ID: 69965a925698665ec4a01e60.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

Recorded evaluation configuration: {"Model backbone":"LLaDA-8B"}. This record does not assert independent reproduction.

### NO-CD

Syntactic Accuracy @1: 74. Result ID: 69965a925698665ec4a01e67.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

Recorded evaluation configuration: {"Model backbone":"Dream-7B"}. This record does not assert independent reproduction.

### NO-CD

Syntactic Accuracy @1: 70.1. Result ID: 69965a925698665ec4a01e6b.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

Recorded evaluation configuration: {"Model backbone":"DiffuC-7B"}. This record does not assert independent reproduction.

### NO-CD

Syntactic Accuracy @1: 66. Result ID: 69965a925698665ec4a01e5f.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

Recorded evaluation configuration: {"Model backbone":"LLaDA-8B"}. This record does not assert independent reproduction.

### NO-CD

Syntactic Accuracy @1: 62.3. Result ID: 69965a925698665ec4a01e63.

[Reporting paper](<https://www.sota2.com/research/paper/69818109a8b01e76cc8e225e.md>)

Recorded evaluation configuration: {"Model backbone":"LLaDA-1.5"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
