LifeSciBench
Expert-authored, artifact-rich free-response tasks that evaluate realistic research judgment across applied life-science workflows.
Evidence before ranking
Map the benchmark. Audit the eval.
A maintained registry of protein, omics, and life-science benchmarks—plus the works, exact models, scopes, tools, graders, and metrics behind every verified evaluation.
35public/metadata benchmark families
195normalized evaluation runs
1573source-located result rows
21/35verified fully open public families
Start with a family
GeneBench-Pro
A research-level agent benchmark of 129 synthetic, multistage computational-biology analyses that require iterative QC, statistical modeling, diagnostics, and decision-relevant judgment.
BioMysteryBench
An agentic bioinformatics benchmark of objective, expert-authored mysteries over anonymized real-world biological data, scored on final answers rather than prescribed analysis paths.
ProteinGym
Versioned deep-mutational-scanning and clinical-variant benchmarks for protein fitness prediction and design in zero-shot and supervised regimes.
CompBioBench
A 100-task agent benchmark of objectively gradable computational-biology problems requiring multi-step reasoning, bespoke code, tools, and real-world external resources.
LAB-Bench
A practical biology-research suite of 2,457 multiple-choice questions across eight broad categories and 31 versioned task files, with public and private contamination-monitoring splits.
Coverage, not a leaderboard
See where evidence clusters—and where it doesn’t.
The same taxonomy powers the explorer, domain pages, and exports. Counts describe registry coverage, not scientific importance.
5 official model-provider works.
Creator sources, official provider reports, peer-reviewed papers, and stable formal preprints are publishable after versioned source review. Private/internal suites are displayed separately and excluded from public coverage rates.
Creator sources, official provider reports, peer-reviewed papers, and stable formal preprints are publishable after versioned source review. Private/internal suites are displayed separately and excluded from public coverage rates.
Domain × capability
Number of benchmark families tagged on both axes.
Accessible data table
| Domain | Capability | Families |
|---|---|---|
| Protein sequence | Prediction | 6 |
| Protein sequence | Design | 5 |
| Protein sequence | Data analysis | 1 |
| Protein sequence | Coding | 0 |
| Protein sequence | Tool use | 1 |
| Protein sequence | Experiment planning | 1 |
| Protein sequence | Scientific reasoning | 3 |
| Protein structure | Prediction | 5 |
| Protein structure | Design | 3 |
| Protein structure | Data analysis | 2 |
| Protein structure | Coding | 2 |
| Protein structure | Tool use | 2 |
| Protein structure | Experiment planning | 0 |
| Protein structure | Scientific reasoning | 3 |
| Protein design | Prediction | 5 |
| Protein design | Design | 5 |
| Protein design | Data analysis | 0 |
| Protein design | Coding | 0 |
| Protein design | Tool use | 0 |
| Protein design | Experiment planning | 0 |
| Protein design | Scientific reasoning | 1 |
| Protein-protein binding | Prediction | 10 |
| Protein-protein binding | Design | 4 |
| Protein-protein binding | Data analysis | 1 |
| Protein-protein binding | Coding | 0 |
| Protein-protein binding | Tool use | 1 |
| Protein-protein binding | Experiment planning | 1 |
| Protein-protein binding | Scientific reasoning | 3 |
| Protein-ligand binding | Prediction | 6 |
| Protein-ligand binding | Design | 1 |
| Protein-ligand binding | Data analysis | 0 |
| Protein-ligand binding | Coding | 0 |
| Protein-ligand binding | Tool use | 0 |
| Protein-ligand binding | Experiment planning | 0 |
| Protein-ligand binding | Scientific reasoning | 1 |
| Genomics | Prediction | 6 |
| Genomics | Design | 1 |
| Genomics | Data analysis | 7 |
| Genomics | Coding | 4 |
| Genomics | Tool use | 7 |
| Genomics | Experiment planning | 1 |
| Genomics | Scientific reasoning | 7 |
| Transcriptomics | Prediction | 3 |
| Transcriptomics | Design | 1 |
| Transcriptomics | Data analysis | 10 |
| Transcriptomics | Coding | 6 |
| Transcriptomics | Tool use | 7 |
| Transcriptomics | Experiment planning | 1 |
| Transcriptomics | Scientific reasoning | 9 |
| Single-cell | Prediction | 0 |
| Single-cell | Design | 0 |
| Single-cell | Data analysis | 9 |
| Single-cell | Coding | 6 |
| Single-cell | Tool use | 6 |
| Single-cell | Experiment planning | 0 |
| Single-cell | Scientific reasoning | 7 |
| Proteomics | Prediction | 0 |
| Proteomics | Design | 0 |
| Proteomics | Data analysis | 3 |
| Proteomics | Coding | 3 |
| Proteomics | Tool use | 3 |
| Proteomics | Experiment planning | 0 |
| Proteomics | Scientific reasoning | 3 |
| Multi-omics | Prediction | 1 |
| Multi-omics | Design | 0 |
| Multi-omics | Data analysis | 2 |
| Multi-omics | Coding | 2 |
| Multi-omics | Tool use | 2 |
| Multi-omics | Experiment planning | 0 |
| Multi-omics | Scientific reasoning | 4 |
Protein and omics coverage
Registry family count by selected scientific domain.
Accessible data table
| Domain | Families |
|---|---|
| Protein sequence | 9 |
| Protein structure | 10 |
| Protein design | 8 |
| Protein-protein binding | 12 |
| Protein-ligand binding | 6 |
| Genomics | 12 |
| Transcriptomics | 13 |
| Single-cell | 10 |
| Proteomics | 3 |
| Multi-omics | 4 |
Scientific object × task family
Distinct benchmark families with eligible task mappings in each cell. Count units are not combined.
Accessible data table
| Scientific object | Task family | Benchmark families |
|---|---|---|
| Protein | Structure prediction | 5 |
| Protein | Design and generation | 4 |
| Protein | Molecular interaction | 8 |
| Protein | Property and function prediction | 5 |
| Protein | Sequence and regulation | 0 |
| Protein | Variant and perturbation effect | 4 |
| Protein | Omics analysis | 0 |
| Protein | Cellular and spatial analysis | 0 |
| Protein | Statistical genetics | 0 |
| Protein | Scientific workflow | 0 |
| Protein | Systems modeling | 0 |
| DNA | Structure prediction | 0 |
| DNA | Design and generation | 1 |
| DNA | Molecular interaction | 2 |
| DNA | Property and function prediction | 0 |
| DNA | Sequence and regulation | 3 |
| DNA | Variant and perturbation effect | 2 |
| DNA | Omics analysis | 0 |
| DNA | Cellular and spatial analysis | 0 |
| DNA | Statistical genetics | 0 |
| DNA | Scientific workflow | 0 |
| DNA | Systems modeling | 0 |
| RNA | Structure prediction | 2 |
| RNA | Design and generation | 1 |
| RNA | Molecular interaction | 2 |
| RNA | Property and function prediction | 2 |
| RNA | Sequence and regulation | 3 |
| RNA | Variant and perturbation effect | 2 |
| RNA | Omics analysis | 0 |
| RNA | Cellular and spatial analysis | 0 |
| RNA | Statistical genetics | 0 |
| RNA | Scientific workflow | 0 |
| RNA | Systems modeling | 0 |
| Small molecule | Structure prediction | 0 |
| Small molecule | Design and generation | 1 |
| Small molecule | Molecular interaction | 6 |
| Small molecule | Property and function prediction | 2 |
| Small molecule | Sequence and regulation | 0 |
| Small molecule | Variant and perturbation effect | 0 |
| Small molecule | Omics analysis | 0 |
| Small molecule | Cellular and spatial analysis | 0 |
| Small molecule | Statistical genetics | 0 |
| Small molecule | Scientific workflow | 0 |
| Small molecule | Systems modeling | 0 |
| Cell | Structure prediction | 0 |
| Cell | Design and generation | 0 |
| Cell | Molecular interaction | 0 |
| Cell | Property and function prediction | 0 |
| Cell | Sequence and regulation | 0 |
| Cell | Variant and perturbation effect | 0 |
| Cell | Omics analysis | 4 |
| Cell | Cellular and spatial analysis | 3 |
| Cell | Statistical genetics | 0 |
| Cell | Scientific workflow | 0 |
| Cell | Systems modeling | 0 |
| Tissue | Structure prediction | 0 |
| Tissue | Design and generation | 0 |
| Tissue | Molecular interaction | 0 |
| Tissue | Property and function prediction | 0 |
| Tissue | Sequence and regulation | 0 |
| Tissue | Variant and perturbation effect | 0 |
| Tissue | Omics analysis | 4 |
| Tissue | Cellular and spatial analysis | 1 |
| Tissue | Statistical genetics | 0 |
| Tissue | Scientific workflow | 0 |
| Tissue | Systems modeling | 0 |
| Organism or population | Structure prediction | 0 |
| Organism or population | Design and generation | 0 |
| Organism or population | Molecular interaction | 0 |
| Organism or population | Property and function prediction | 0 |
| Organism or population | Sequence and regulation | 0 |
| Organism or population | Variant and perturbation effect | 0 |
| Organism or population | Omics analysis | 0 |
| Organism or population | Cellular and spatial analysis | 0 |
| Organism or population | Statistical genetics | 0 |
| Organism or population | Scientific workflow | 0 |
| Organism or population | Systems modeling | 0 |
| Microbial community | Structure prediction | 0 |
| Microbial community | Design and generation | 0 |
| Microbial community | Molecular interaction | 0 |
| Microbial community | Property and function prediction | 0 |
| Microbial community | Sequence and regulation | 0 |
| Microbial community | Variant and perturbation effect | 0 |
| Microbial community | Omics analysis | 0 |
| Microbial community | Cellular and spatial analysis | 0 |
| Microbial community | Statistical genetics | 0 |
| Microbial community | Scientific workflow | 0 |
| Microbial community | Systems modeling | 0 |
| Omics profile | Structure prediction | 0 |
| Omics profile | Design and generation | 0 |
| Omics profile | Molecular interaction | 0 |
| Omics profile | Property and function prediction | 0 |
| Omics profile | Sequence and regulation | 0 |
| Omics profile | Variant and perturbation effect | 0 |
| Omics profile | Omics analysis | 5 |
| Omics profile | Cellular and spatial analysis | 3 |
| Omics profile | Statistical genetics | 0 |
| Omics profile | Scientific workflow | 5 |
| Omics profile | Systems modeling | 0 |
| Experimental system | Structure prediction | 0 |
| Experimental system | Design and generation | 0 |
| Experimental system | Molecular interaction | 0 |
| Experimental system | Property and function prediction | 0 |
| Experimental system | Sequence and regulation | 0 |
| Experimental system | Variant and perturbation effect | 0 |
| Experimental system | Omics analysis | 0 |
| Experimental system | Cellular and spatial analysis | 0 |
| Experimental system | Statistical genetics | 0 |
| Experimental system | Scientific workflow | 6 |
| Experimental system | Systems modeling | 1 |
Explore all Scientific Tasks, aliases, counts, and coverage gaps →
Benchmark release timeline
Initial public release date for benchmark families with a verified, non-provisional kind classification.
Accessible data table
| Benchmark | Release date | Kind |
|---|---|---|
| AB-Bind | Fri Nov 06 2015 00:00:00 GMT+0000 (Coordinated Universal Time) | dataset |
| AbBiBench | Fri May 23 2025 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| ATOM3D | Mon Dec 07 2020 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| BEACON | Fri Jun 14 2024 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| benchmark-dual human CRISPR-Cas9 library | Wed Feb 26 2025 00:00:00 GMT+0000 (Coordinated Universal Time) | dataset |
| benchmark human CRISPR-Cas9 library | Wed Feb 26 2025 00:00:00 GMT+0000 (Coordinated Universal Time) | dataset |
| Biology-Instructions | Thu Dec 26 2024 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| BioMysteryBench | Wed Apr 29 2026 00:00:00 GMT+0000 (Coordinated Universal Time) | agentic-eval |
| BioSecBench-Surveillance | Tue Jul 21 2026 00:00:00 GMT+0000 (Coordinated Universal Time) | agentic-eval |
| BixBench | Fri Feb 28 2025 00:00:00 GMT+0000 (Coordinated Universal Time) | agentic-eval |
| BLADE | Mon Aug 19 2024 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| CaM benchmark | Thu Jul 11 2024 00:00:00 GMT+0000 (Coordinated Universal Time) | dataset |
| CAMEO | Sun Jan 01 2012 00:00:00 GMT+0000 (Coordinated Universal Time) | competition |
| CASP | Sat Jan 01 1994 00:00:00 GMT+0000 (Coordinated Universal Time) | competition |
| CompBioBench | Mon Apr 06 2026 00:00:00 GMT+0000 (Coordinated Universal Time) | agentic-eval |
| crafted experiments | Tue Jan 07 2025 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| FLIP | Mon Oct 11 2021 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| GeneBench-Pro | Tue Jun 30 2026 00:00:00 GMT+0000 (Coordinated Universal Time) | agentic-eval |
| Genomic Benchmarks | Mon May 01 2023 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| GuacaMol | Fri Nov 23 2018 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| LAB-Bench | Sun Jul 14 2024 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| MoleculeNet | Thu Mar 02 2017 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| PapD benchmark | Thu Jul 11 2024 00:00:00 GMT+0000 (Coordinated Universal Time) | dataset |
| PPB-Affinity | Tue Dec 03 2024 00:00:00 GMT+0000 (Coordinated Universal Time) | dataset |
| ProteinGym | Fri Dec 08 2023 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| ProteinLMBench | Mon Apr 29 2024 00:00:00 GMT+0000 (Coordinated Universal Time) | dataset |
| RfaH benchmark | Thu Jul 11 2024 00:00:00 GMT+0000 (Coordinated Universal Time) | dataset |
| scBench | Mon Feb 09 2026 00:00:00 GMT+0000 (Coordinated Universal Time) | agentic-eval |
| scIB | Thu Dec 23 2021 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| SCIGYM | Fri May 16 2025 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| Single-cell Omics Arena | Mon Nov 24 2025 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| SpatialBench | Fri Dec 26 2025 00:00:00 GMT+0000 (Coordinated Universal Time) | agentic-eval |
| TAPE | Wed Jun 19 2019 00:00:00 GMT+0000 (Coordinated Universal Time) | suite |
| VirBench | Tue May 20 2025 00:00:00 GMT+0000 (Coordinated Universal Time) | agentic-eval |
Registry activity
New or re-verified evidence in the last 30 / 90 days
34 / 34papers and works
10 / 10benchmark-use relations
190 / 190normalized runs
2026-07-22registry as-of date
Counts use versioned review timestamps, not paper publication dates. Candidate issues are excluded until owner-reviewed and merged.
Machine-readable by default
One normalized object, every public surface.
The website, JSON, and CSV exports are built from the same verified YAML records. Draft entries appear in previews, never in production exports.
Public interfaces
Nucleic-acid benchmark results
Browse protocol-aligned baselines, original-table bests, official-board leaders, and evidence across 47 benchmarks and 58 tasks. No cross-benchmark leaderboard.