Anthropic Scientific Figure Interpretation Eval
Private Anthropic evaluation direction for scientific figure interpretation, reported only through a model-trend chart with no task count or released examples.
Scientific domain
Molecular and cell biology
Cellular mechanisms
分子与细胞生物学
25 registered benchmarks and tracks
BEACON
A 13-task RNA representation benchmark covering structure, function, processing, modification, translation, degradation, programmable switches, and CRISPR activity.
benchmark-dual human CRISPR-Cas9 library
A human CRISPR-Cas9 paired-guide library created to compare dual- and single-targeting strategies in loss-of-function screens.
benchmark human CRISPR-Cas9 library
A human CRISPR-Cas9 guide-RNA library assembled to compare single-targeting library performance in loss-of-function screens.
Biology-Instructions Enhancer-Promoter Interaction Prediction
Binary interaction prediction for enhancer and promoter DNA sequences, evaluated with Matthews correlation coefficient.
Biology-Instructions Programmable RNA Switches
Three-output regression of ON, OFF, and ON/OFF programmable RNA-switch values, aggregated as their mean squared Pearson correlation.
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.
LAB-Bench CloningScenarios
Human-hard, multi-step multiple-choice scenarios involving plasmids, DNA fragments, enzymes, and molecular-cloning workflows.
LAB-Bench FigQA
Multiple-choice interpretation and multi-element reasoning over scientific figures shown without captions or paper context.
LAB-Bench ProtocolQA
Troubleshoots intentionally modified published biological protocols by selecting steps that would repair the stated outcome.
LAB-Bench SeqQA
Sequence-comprehension and manipulation category spanning 15 formal tasks involving PCR, restriction digestion, ORFs, translation, GC content, and DNA–protein relationships.
LAB-Bench SeqQA — Gene-to-restriction primers
Selects restriction-cloning primers from a named gene and enzyme pair.
LAB-Bench SeqQA — Gene-to-Gibson primers (HindIII)
Selects primers for Gibson assembly into a HindIII-linearized vector.
LAB-Bench SeqQA — Gene-to-Gibson primers (SmaI)
Selects primers for Gibson assembly into a SmaI-linearized vector.
LAB-Bench SeqQA — Primers-to-restriction enzymes
Infers restriction enzymes from a gene name and primer pair.
LAB-Bench SeqQA — Amplicon length to primers
Selects primers that produce a requested amplicon length from a DNA template.
LAB-Bench SeqQA — Primers to amplicon length
Calculates expected amplicon length from a primer pair and DNA template.
LAB-Bench SeqQA — Sequence-to-restriction primers
Selects restriction-cloning primers from an explicit gene sequence and enzyme pair.
LAB-Bench SeqQA — Amplicon sequence to primers
Selects primers that produce a requested amplicon sequence from a DNA template.
LAB-Bench SeqQA — Restriction-fragment lengths
Calculates fragment lengths after restriction digestion of a DNA sequence.
LAB-Bench SeqQA — Restriction-fragment count
Calculates the number of fragments after restriction digestion of a DNA sequence.
LifeSciBench
Expert-authored, artifact-rich free-response tasks that evaluate realistic research judgment across applied life-science workflows.
SCIGYM
An agentic systems-biology suite in which language models iteratively perturb simulated SBML systems, analyze time-series observations in Python, and reconstruct hidden biological reactions.
SCIGYM Large
The formally released SCIGYM track containing the 213 systems not included in the creator paper's model evaluation, with systems reaching up to 400 reactions.
SCIGYM Small
The formally released and creator-evaluated SCIGYM track containing biological systems with fewer than ten reactions.
Capability coverage
Registry records tagged Molecular and cell biology, counted by capability.
Accessible data table
| Capability | Records |
|---|---|
| Knowledge | 1 |
| Evidence synthesis | 4 |
| Retrieval | 2 |
| Prediction | 7 |
| Classification | 3 |
| Regression | 2 |
| Design | 10 |
| Generation | 1 |
| Optimization | 1 |
| Data analysis | 11 |
| Coding | 3 |
| Tool use | 5 |
| Experiment planning | 6 |
| Troubleshooting | 4 |
| Scientific reasoning | 18 |
| Scientific communication | 1 |
Scientific Tasks in this domain
Task mappings are evidence-backed and may be partial for mixed suites.
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.