Back to results

CompBioBench item

covid-patient-q1

Question

covid.patients.q1.h5ad contains PBMC scRNA-seq data from 33 donors, who are a mix of healthy and COVID patients. However, the patient metadata is missing. Identify which donors are healthy and which have COVID based on the gene expression data. Provide a single string containing comma-separated donor IDs for the healthy patients (numerically sorted), e.g."1,2,5".

Inferred answer
1,5,11,12,17,21,24,27,29,32Uniquely inferred from submitted scores.
Domain
Single-cell
Question style
Metadata Recovery

An answer can be uniquely inferred for 54 of 100 questions from submitted scores. Green: correct · Red: incorrect. For questions without a uniquely determined answer: Blue: matches prediction · Orange: differs from prediction. These are inferred answer checks, not official per-question grades.

Galaxy

12/12 terminal item results

UDT = user-defined tool: custom code executed by Galaxy. Route badges summarize analysis tools, excluding uploads and automatic conversions. How usage is counted.

Codex GPT-5.5

3 replicates
Codex GPT-5.5

Replicate 1

CorrectUDT-only
Answer1,5,11,12,17,21,24,27,29,32
Workflow
COVID H5AD donor diagnosticCOVID H5AD direct donor diagnosticCOVID H5AD normalized signaturesCOVID H5AD inventoryCOVID pseudobulk donor cluster2 more tools
Total tokens
1.6M
Output
21.8K
Codex GPT-5.5

Replicate 2

CorrectUDT-only
Answer1,5,11,12,17,21,24,27,29,32
Workflow
COVID H5AD diagnosticCOVID H5AD diagnostic v2COVID H5AD h5py diagnosticCOVID H5AD signature analysisCOVID H5AD signature analysis v21 more tools
Total tokens
912.6K
Output
16.4K
Codex GPT-5.5

Replicate 3

CorrectUDT + standard tools
Answer1,5,11,12,17,21,24,27,29,32
Workflow
COVID h5ad donor diagnosticCOVID h5ad donor diagnostic v2COVID h5ad donor diagnostic v3COVID h5ad donor diagnostic v4Anndata Inspect3 more tools
Total tokens
3.9M
Output
22.6K

Codex GPT-5.6 Sol

3 replicates
Codex GPT-5.6 Sol

Replicate 1

CorrectUDT-only
Answer1,5,11,12,17,21,24,27,29,32
Workflow
COVID donor structural and marker diagnosticAnnData
Total tokens
6.6M
Output
20.7K
Codex GPT-5.6 Sol

Replicate 2

CorrectUDT + standard tools
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Anndata InspectCOVID donor expression diagnosticCOVID donor cell-stratified validationCOVID donor final expression call
Total tokens
5.5M
Output
22.0K
Codex GPT-5.6 Sol

Replicate 3

CorrectUDT + standard tools
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Anndata InspectCOVID donor expression diagnosticSelect columnsCOVID donor classification validationExtract validated healthy donor IDs
Total tokens
3.8M
Output
15.8K

Codex + DeepSeek-v4-pro-0813

3 replicates
Codex + DeepSeek-v4-pro-0813

Replicate 1

RerunIncorrectUDT + standard tools
Answer1,2,4,5,6,7,8,9,10,11,12,15,16,17,19,20,21,22,23,24,25,27,28,29,32
Workflow
anndata inspect → anndata export → User-defined tool
Total tokens
23.1M
Output
58.6K
Codex + DeepSeek-v4-pro-0813

Replicate 2

RerunIncorrectUDT + standard tools
Answer1,5,6,8,10,11,12,15,17,19,21,22,23,24,25,27,29,32
Workflow
anndata inspect → scanpy inspect → scanpy normalize → decoupler pseudobulk → datamash ops → sklearn numeric clustering → sklearn pca → User-defined tool
Total tokens
18.3M
Output
67.5K
Codex + DeepSeek-v4-pro-0813

Replicate 3

IncorrectUDT + standard tools
Answer1,2,4,5,6,7,9,10,11,12,15,16,17,20,21,22,23,24,25,27,28,29,32
Workflow
Anndata InspectAnndata ExportCOVID donor metricsUDT smokeUDT smoke scanpy7 more tools
Total tokens
74.7M
Output
173.7K

Codex GPT-5.6 Luna

3 replicates
Codex GPT-5.6 Luna

Replicate 1

IncorrectUDT + standard tools
Answer1,2,4,5,6,7,8,9,10,11,12,15,16,17,19,20,21,22,23,24,25,27,28,29,32
Workflow
Anndata InspectFilter TabularH5AD Runtime ProbeCOVID Donor Expression ExplorationH5AD Metadata Probe4 more tools
Total tokens
13.1M
Output
48.3K
Codex GPT-5.6 Luna

Replicate 2

CorrectUDT + standard tools
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Anndata InspectCOVID donor H5AD runtime probeCOVID H5AD structure probeCOVID donor expression diagnosticCOVID donor expression validation1 more tools
Total tokens
3.9M
Output
32.5K
Codex GPT-5.6 Luna

Replicate 3

IncorrectUDT + standard tools
Answer1,2,4,5,6,7,8,9,10,11,12,15,16,17,19,20,21,22,23,24,25,27,28,29,32
Workflow
COVID donor response diagnosticAnnData
Total tokens
7.3M
Output
49.8K

Anycode

13/13 terminal item results

Codex GPT-5.5

3 replicates
Codex GPT-5.5

Replicate 1

Correct
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Local codePythonAnnDataScanpySciPy
Codex GPT-5.5

Replicate 2

Correct
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Local codePythonAnnDataScanpySciPy
Codex GPT-5.5

Replicate 3

Correct
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Local codePythonAnnDataScanpySciPy

Codex GPT-5.6 Sol

3 replicates
Codex GPT-5.6 Sol

Replicate 1

RerunCorrect
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Open-ended code
Total tokens
997.3K
Output
11.1K
Codex GPT-5.6 Sol

Replicate 2

Correct
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Local codePythonAnnDataScanpySciPy
Codex GPT-5.6 Sol

Replicate 3

Correct
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Local codePythonAnnDataScanpySciPy

Codex + DeepSeek-v4-pro-0813

3 replicates
Codex + DeepSeek-v4-pro-0813

Replicate 1

Incorrect
Answer1,5,6,10,11,12,16,17,21,23,24,27,29,32
Workflow
Local codePythonAnnDataScanpySciPy1 more tools
Total tokens
57.5M
Output
231.1K
Codex + DeepSeek-v4-pro-0813

Replicate 2

RerunCorrect
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Open-ended code
Total tokens
11.1M
Output
52.1K
Codex + DeepSeek-v4-pro-0813

Replicate 3

Correct
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Local codePythonAnnDataScanpySciPy
Total tokens
21.4M
Output
90.3K

Codex GPT-5.6 Luna

3 replicates
Codex GPT-5.6 Luna

Replicate 1

RerunCorrect
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Open-ended code
Total tokens
1.2M
Output
19.5K
Codex GPT-5.6 Luna

Replicate 2

Correct
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Local codePythonAnnDataScanpySciPy
Codex GPT-5.6 Luna

Replicate 3

Incorrect
Answer1,2,4,5,6,7,8,9,10,11,12,13,15,16,17,19,20,21,22,23,24,25,27,28,29,30,31,32
Workflow
Local codePythonAnnDataScanpySciPy

Codex GPT-6 Astra

1 replicate
Codex GPT-6 Astra

Replicate 1

Correct
Answer1,5,11,12,17,21,24,27,29,32
Workflow
Local codePythonAnnDataScanpySciPy
Total tokens
801.7K
Output
7.3K