Scientific domain

Protein science

Protein sequence

蛋白质科学

9 registered benchmarks and tracks

suite0 runs

AbBiBench

A framework using antibody–antigen complexes to evaluate affinity prediction and antibody redesign.

PredictionDesignGenerationOptimization
suite0 runs

Anthropic Key Life Sciences Evals

An Anthropic private internal suite reported only through an official accuracy chart covering scientific figure interpretation, computational biology, and protein understanding.

KnowledgeEvidence synthesisData analysisScientific reasoning
track1 runs

Anthropic Protein Understanding Eval

Private Anthropic protein-understanding evaluation direction reported only through a model-trend chart, without public tasks, counts, or protocol details.

KnowledgeScientific reasoning
dataset0 runs

CaM benchmark

A multistate protein sequence-design benchmark spanning CaM conformations and binding modes.

DesignOptimization
agentic-eval1 runs

LifeSciBench

Expert-authored, artifact-rich free-response tasks that evaluate realistic research judgment across applied life-science workflows.

Evidence synthesisRetrievalDesignGeneration
dataset0 runs

PapD benchmark

A multistate protein sequence-design benchmark targeting the multispecific PapD binding interface.

DesignOptimization
dataset1 runs

ProteinLMBench

A creator-curated set of 944 protein-science multiple-choice questions with answer explanations, generated from research literature and released for evaluating text LLM protein understanding.

KnowledgeScientific reasoningPrediction
dataset0 runs

RfaH benchmark

A multistate protein sequence-design benchmark using the fold-switching conformations of RfaH.

DesignOptimization

Capability coverage

Registry records tagged Protein science, counted by capability.

CSV ↓
Accessible data table
CapabilityRecords
Knowledge4
Evidence synthesis2
Retrieval2
Prediction3
Design5
Generation2
Optimization5
Data analysis2
Experiment planning1
Troubleshooting1
Scientific reasoning4
Scientific communication1
Coverage-gap reading: zero counts indicate a gap in this registry, not proof that no benchmark exists. Propose a primary source through the contribution form.