agentic-eval · audited · verified 2026-07-21

CompBioBench

A 100-task agent benchmark of objectively gradable computational-biology problems requiring multi-step reasoning, bespoke code, tools, and real-world external resources.

+4 more

Task, access, and protocol audit: v1 contains 100 tasks. The public TSV partitions them into 8 domains and 5 question styles; only one Structure task uses a protein PDB, while standalone binding counts are Not reported. Tasks and inputs are open, but the answer key and grader backend are private. Creator results are split by agent, effort, repeats, timeout policy, and the 17-task hardest subset.

Benchmark definition

What is counted

Version: v1
Total: 100 (v1 independent computational-biology tasks)
Task formats: agentic computational analysis; exact single-line answer
Capabilities: Data analysisCodingTool useRetrievalScientific reasoning
Modalities: TextTableDNA or RNA sequenceProtein sequence3D structureRaw omicsDatabaseWebCode

Version history

Version	Status	Release / as-of	Total	Formal tracks
v1 `compbiobench-v1`	current	2026-04-06	100 (v1 independent computational-biology tasks)	None registered

Tracks and subsets

ID	Count	Basis	Partition?	Notes
Domain — Epigenomics `compbiobench-domain-epigenomics`	20	tasks in the domain partition	No	Official v1 TSV domain label; partition members are non-exhaustive here because other independent partitions are also registered.
Domain — Genomics `compbiobench-domain-genomics`	20	tasks in the domain partition	No	Official v1 TSV domain label.
Domain — Machine Learning `compbiobench-domain-machine-learning`	7	tasks in the domain partition	No	Official v1 TSV domain label.
Domain — Population Genetics `compbiobench-domain-population-genetics`	12	tasks in the domain partition	No	Official v1 TSV domain label.
Domain — Single-cell `compbiobench-domain-single-cell`	21	tasks in the domain partition	No	Official v1 TSV domain label.
Domain — Spatial `compbiobench-domain-spatial`	2	tasks in the domain partition	No	Official v1 TSV domain label.
Domain — Structure `compbiobench-domain-structure`	1	tasks in the domain partition	No	The sole Structure task asks which uppercase letter a supplied PDB protein resembles across projections.
Domain — Transcriptomics `compbiobench-domain-transcriptomics`	17	tasks in the domain partition	No	Official v1 TSV domain label.
Question style — Metadata Recovery `compbiobench-style-metadata-recovery`	27	tasks in the question-style partition	No	Official v1 TSV question_style label.
Question style — Retrieval `compbiobench-style-retrieval`	17	tasks in the question-style partition	No	Official v1 TSV question_style label.
Question style — Routine Analysis `compbiobench-style-routine-analysis`	22	tasks in the question-style partition	No	Official v1 TSV question_style label.
Question style — Synthetic/Augmented Data `compbiobench-style-synthetic-augmented`	26	tasks in the question-style partition	No	Official v1 TSV question_style label.
Question style — Tooling `compbiobench-style-tooling`	8	tasks in the question-style partition	No	Official v1 TSV question_style label.
Contributor difficulty — Level 1 `compbiobench-difficulty-level-1`	17	tasks in the paper difficulty partition	No	Contributor ratings are approximate and were not extensively calibrated.
Contributor difficulty — Level 2 `compbiobench-difficulty-level-2`	26	tasks in the paper difficulty partition	No	Contributor ratings are approximate and were not extensively calibrated.
Contributor difficulty — Level 3 `compbiobench-difficulty-level-3`	40	tasks in the paper difficulty partition	No	Contributor ratings are approximate and were not extensively calibrated.
Contributor difficulty — Levels 4–5 `compbiobench-difficulty-levels-4-5`	17	tasks in the grouped hardest subset	No	The paper groups Levels 4 and 5 for its hardest-subset analysis.
Internet required `compbiobench-internet-required`	78	tasks whose v1 TSV internet_required field is True	No	Public v1 TSV count.
Internet not required `compbiobench-internet-not-required`	22	tasks whose v1 TSV internet_required field is False	No	Public v1 TSV count.

Scientific Task Atlas

Scientific task classification

partial for v1. Official domains are not an exhaustive scientific-task taxonomy; the runner establishes the end-to-end workflow.

Scientific task	Coverage	Count	Mapping	Evidence
End-to-end computational analysis	explicitly-in-scope	100 tasks v1 independent computational-biology tasks	official-taxonomy high confidence	`compbiobench-evidence-counts` `compbiobench-evidence-runner-license` Tasks require code, tools, and external resources.
Omics and cellular analysis	explicitly-in-scope	Not reported v1 tasks in official genomics, transcriptomics, epigenetics, single-cell, and spatial domains.	official-taxonomy high confidence	`compbiobench-evidence-taxonomy` The official domains establish broad omics coverage but do not provide an exhaustive Scientific Task breakdown.

Scientific coverage notes

Domain	Coverage	Count	Interpretation
Protein structure	observed	1	The official v1 TSV contains one Structure-domain PDB projection task; this is not a protein-folding or binding benchmark.
Protein-protein binding	unknown	Not reported	No standalone protein-protein binding category or count is reported.
Protein-ligand binding	unknown	Not reported	No standalone protein-ligand binding category or count is reported.

Evaluation registry

Works and run settings

A setting change—scope, prompt, tools, budget, grader, or repeats—creates a separate run. Charts never cross a comparability group.

Evaluation run

compbiobench-codex-hardest

From Agentic systems are adept at solving well-scoped, verifiable problems in computational biology

Evaluated models / systems: Codex CLI (GPT-5.4)

Scopesubset · n=17

Shotszero-shot

Turnsmulti-turn agent trajectory

System prompt publicNo

Reasoning / effortxhigh

Browserweb access through agent environment

InternetYes

DatabasesYes

Code executionYes

ContainerNo

External toolsYes

Token budgetNot reported

Time / cost budget120 minutes initially; timeout rerun once with 240 minutes

TemperatureNot reported

SeedNot reported

Repeats3

Graderwhitespace-stripped exact string match with occasional manual format-only correction · human review: yes

Statisticsaccuracy averaged over three runs within the 17-task subset

Contaminationbenchmark construction reduces lookup shortcuts

Metrics, results, and full protocol

Metrics

Metric	Kind / baseline	Unit	Aggregation	Threshold / tolerance
Accuracy — difficulty Levels 4–5	absolute	percent	mean problem accuracy over three runs	whitespace-stripped exact match

Results

Model	Metric	Value	n
Codex CLI (GPT-5.4)	Accuracy — difficulty Levels 4–5	59 percent Supplementary Figure 2 label; average across three runs.	17

Evidence

page: PDF pp. 13–14, Methods (Exact Codex configuration and common agent protocol.) — supports /protocol
figure: PDF p. 11, Supplementary Figure 2B and caption (Levels 4 and 5 grouped; 17 tasks; Codex label 59; three-run average.) — supports /scope, /metrics, /results

Evaluation run

compbiobench-creator-full

From Agentic systems are adept at solving well-scoped, verifiable problems in computational biology

Evaluated models / systems: Codex CLI (GPT-5.4)

Scopefull · n=100

Shotszero-shot

Turnsmulti-turn agent trajectory

System prompt publicNo

Reasoning / effortxhigh

Browserweb access through the agent environment

InternetYes

DatabasesYes

Code executionYes

ContainerNo

External toolsYes

Token budgetNot reported

Time / cost budget120 minutes initially; timed-out questions rerun clean once with 240 minutes

TemperatureNot reported

SeedNot reported

Repeats3

Graderwhitespace-stripped exact string match with occasional manual format-only correction · human review: yes

Statisticsmean over three independent runs; consistency reported as correct in all three and correct at least once

Contaminationsynthetic/augmented inputs and scrubbed identifiers reduce lookup shortcuts

Metrics, results, and full protocol

Metrics

Metric	Kind / baseline	Unit	Aggregation	Threshold / tolerance
Accuracy	absolute	percent	mean problem accuracy over three runs	whitespace-stripped exact match
Wall-clock time per question	absolute	seconds	mean over questions and runs after 7200-second display clipping	Not reported
Cost per question	absolute	USD	mean over questions and runs after USD-10 display clipping	Not reported

Results

Model	Metric	Value	n
Codex CLI (GPT-5.4)	Accuracy	83.3 percent Mean of three full runs; 73% solved on all three and 92% at least once.	100
Codex CLI (GPT-5.4)	Wall-clock time per question	679 seconds Printed Figure 2 label; mean across three runs.	100
Codex CLI (GPT-5.4)	Cost per question	1 USD Printed Figure 2 label; mean across three runs.	100

Evidence

page: PDF pp. 13–14, Methods: agent execution and model configurations (Reports public wrapper prompt, internet/code/tool access, Conda isolation, 120/240-minute policy, Codex CLI v0.115.0, gpt-5.4, xhigh, and three runs.) — supports /benchmark_version, /scope, /protocol
figure: PDF pp. 3–4, Figure 2A–C and Results (Printed labels report 83.3% accuracy, 679.0 seconds, and USD 1.0; text reports three-run consistency.) — supports /metrics, /results

Evaluation run

compbiobench-haiku-full

From Agentic systems are adept at solving well-scoped, verifiable problems in computational biology

Evaluated models / systems: Claude Code (Haiku 4.5)

Scopefull · n=100

Shotszero-shot

Turnsmulti-turn agent trajectory

System prompt publicNo

Reasoning / effortunsupported

Browserweb access through the agent environment

InternetYes

DatabasesYes

Code executionYes

ContainerNo

External toolsYes

Token budgetNot reported

Time / cost budget120 minutes; no 240-minute timeout rerun

TemperatureNot reported

SeedNot reported

Repeats1

Graderwhitespace-stripped exact string match with occasional manual format-only correction · human review: yes

Statisticssingle full-benchmark run

Contaminationsynthetic/augmented inputs and scrubbed identifiers reduce lookup shortcuts

Metrics, results, and full protocol

Metrics

Metric	Kind / baseline	Unit	Aggregation	Threshold / tolerance
Accuracy	absolute	percent	problem-weighted mean	whitespace-stripped exact match
Wall-clock time per question	absolute	seconds	mean after 7200-second display clipping	Not reported
Cost per question	absolute	USD	mean after USD-10 display clipping	Not reported

Results

Model	Metric	Value	n
Claude Code (Haiku 4.5)	Accuracy	34 percent One full run.	100
Claude Code (Haiku 4.5)	Wall-clock time per question	809.3 seconds Printed Figure 2 label.	100
Claude Code (Haiku 4.5)	Cost per question	0.3 USD Printed Figure 2 label.	100

Evidence

page: PDF pp. 13–14, Methods (Reports Claude Code v2.1.87, dated Haiku identifier, unsupported effort, one run, and no timeout rerun.) — supports /benchmark_version, /scope, /protocol
figure: PDF pp. 3–4, Figure 2A–C (Printed labels report 34.0%, 809.3 seconds, and USD 0.3.) — supports /metrics, /results

Evaluation run

compbiobench-haiku-hardest

From Agentic systems are adept at solving well-scoped, verifiable problems in computational biology

Evaluated models / systems: Claude Code (Haiku 4.5)

Scopesubset · n=17

Shotszero-shot

Turnsmulti-turn agent trajectory

System prompt publicNo

Reasoning / effortunsupported

Browserweb access through agent environment

InternetYes

DatabasesYes

Code executionYes

ContainerNo

External toolsYes

Token budgetNot reported

Time / cost budget120 minutes; no 240-minute timeout rerun

TemperatureNot reported

SeedNot reported

Repeats1

Graderwhitespace-stripped exact string match with occasional manual format-only correction · human review: yes

Statisticsproblem-weighted accuracy in one run over the 17-task subset

Contaminationbenchmark construction reduces lookup shortcuts

Metrics, results, and full protocol

Metrics

Metric	Kind / baseline	Unit	Aggregation	Threshold / tolerance
Accuracy — difficulty Levels 4–5	absolute	percent	problem-weighted mean	whitespace-stripped exact match

Results

Model	Metric	Value	n
Claude Code (Haiku 4.5)	Accuracy — difficulty Levels 4–5	12 percent Supplementary Figure 2 label; one run.	17

Evidence

page: PDF pp. 13–14, Methods (Exact Haiku configuration and timeout exception.) — supports /protocol
figure: PDF p. 11, Supplementary Figure 2B and caption (Levels 4 and 5 grouped; 17 tasks; Haiku label 12.) — supports /scope, /metrics, /results

Evaluation run

compbiobench-nonagentic-baselines

From Agentic systems are adept at solving well-scoped, verifiable problems in computational biology

Evaluated models / systems: ChatGPT 5.2, Claude Opus 4.6

Scopefull · n=100

Shotszero-shot

Turnssingle-turn API call

System prompt publicNot reported

Reasoning / effortdefault API parameters

BrowserNo

InternetNo

DatabasesNo

Code executionNo

ContainerNo

External toolsNo

Token budgetNot reported

Time / cost budgetNot reported

Temperaturedefault

SeedNot reported

Repeats3

Graderwhitespace-stripped exact string match · human review: not reported

Statisticsmean accuracy over three calls per question

Contaminationbenchmark construction reduces lookup shortcuts

Metrics, results, and full protocol

Metrics

Metric	Kind / baseline	Unit	Aggregation	Threshold / tolerance
Accuracy	absolute	percent	mean over three calls per problem	whitespace-stripped exact match

Results

Model	Metric	Value	n
ChatGPT 5.2	Accuracy	5.3 percent ChatGPT 5.2 non-agentic API baseline; three calls per question.	100
Claude Opus 4.6	Accuracy	3.7 percent Claude Opus 4.6 non-agentic API baseline; three calls per question.	100

Evidence

page: PDF p. 13, Methods: LLM-only baselines (Default API parameters, three calls per question, no files, public prompt, and full benchmark.) — supports /benchmark_version, /scope, /protocol
figure: PDF pp. 3–4, Figure 2A and Results (Figure labels and text report ChatGPT 5.2 at 5.3% and Claude Opus 4.6 at 3.7%.) — supports /metrics, /results

Evaluation run

compbiobench-opus-full

From Agentic systems are adept at solving well-scoped, verifiable problems in computational biology

Evaluated models / systems: Claude Code (Opus 4.6)

Scopefull · n=100

Shotszero-shot

Turnsmulti-turn agent trajectory

System prompt publicNo

Reasoning / effortmax

Browserweb access through the agent environment

InternetYes

DatabasesYes

Code executionYes

ContainerNo

External toolsYes

Token budgetNot reported

Time / cost budget120 minutes initially; timed-out questions rerun clean once with 240 minutes

TemperatureNot reported

SeedNot reported

Repeats3

Graderwhitespace-stripped exact string match with occasional manual format-only correction · human review: yes

Statisticsmean over three independent runs; consistency reported as correct in all three and at least once

Contaminationsynthetic/augmented inputs and scrubbed identifiers reduce lookup shortcuts

Metrics, results, and full protocol

Metrics

Metric	Kind / baseline	Unit	Aggregation	Threshold / tolerance
Accuracy	absolute	percent	mean problem accuracy over three runs	whitespace-stripped exact match
Wall-clock time per question	absolute	seconds	mean after 7200-second display clipping	Not reported
Cost per question	absolute	USD	mean after USD-10 display clipping	Not reported

Results

Model	Metric	Value	n
Claude Code (Opus 4.6)	Accuracy	81 percent Mean of three full runs; 73% solved all three and 86% at least once.	100
Claude Code (Opus 4.6)	Wall-clock time per question	1101 seconds Printed Figure 2 label.	100
Claude Code (Opus 4.6)	Cost per question	1.7 USD Printed Figure 2 label.	100

Evidence

page: PDF pp. 13–14, Methods: agent execution and model configurations (Reports Claude Code v2.1.87, claude-opus-4-6 1M, max effort, three runs, tools, and timeout policy.) — supports /benchmark_version, /scope, /protocol
figure: PDF pp. 3–4, Figure 2A–C and Results (Reports 81.0% accuracy, 1101.0 seconds, USD 1.7, and consistency.) — supports /metrics, /results

Evaluation run

compbiobench-opus-hardest

From Agentic systems are adept at solving well-scoped, verifiable problems in computational biology

Evaluated models / systems: Claude Code (Opus 4.6)

Scopesubset · n=17

Shotszero-shot

Turnsmulti-turn agent trajectory

System prompt publicNo

Reasoning / effortmax

Browserweb access through agent environment

InternetYes

DatabasesYes

Code executionYes

ContainerNo

External toolsYes

Token budgetNot reported

Time / cost budget120 minutes initially; timeout rerun once with 240 minutes

TemperatureNot reported

SeedNot reported

Repeats3

Graderwhitespace-stripped exact string match with occasional manual format-only correction · human review: yes

Statisticsaccuracy averaged over three runs within the 17-task subset

Contaminationbenchmark construction reduces lookup shortcuts

Metrics, results, and full protocol

Metrics

Metric	Kind / baseline	Unit	Aggregation	Threshold / tolerance
Accuracy — difficulty Levels 4–5	absolute	percent	mean problem accuracy over three runs	whitespace-stripped exact match

Results

Model	Metric	Value	n
Claude Code (Opus 4.6)	Accuracy — difficulty Levels 4–5	69 percent Supplementary Figure 2 label; average across three runs.	17

Evidence

page: PDF pp. 13–14, Methods (Exact Claude Code Opus configuration and common protocol.) — supports /protocol
figure: PDF p. 11, Supplementary Figure 2B and caption (Levels 4 and 5 grouped; 17 tasks; Opus label 69; three-run average.) — supports /scope, /metrics, /results

Evaluation run

compbiobench-sonnet-full

From Agentic systems are adept at solving well-scoped, verifiable problems in computational biology

Evaluated models / systems: Claude Code (Sonnet 4.6)

Scopefull · n=100

Shotszero-shot

Turnsmulti-turn agent trajectory

System prompt publicNo

Reasoning / efforthigh

Browserweb access through the agent environment

InternetYes

DatabasesYes

Code executionYes

ContainerNo

External toolsYes

Token budgetNot reported

Time / cost budget120 minutes initially; timed-out questions rerun clean once with 240 minutes

TemperatureNot reported

SeedNot reported

Repeats1

Graderwhitespace-stripped exact string match with occasional manual format-only correction · human review: yes

Statisticssingle full-benchmark run

Contaminationsynthetic/augmented inputs and scrubbed identifiers reduce lookup shortcuts

Metrics, results, and full protocol

Metrics

Metric	Kind / baseline	Unit	Aggregation	Threshold / tolerance
Accuracy	absolute	percent	problem-weighted mean	whitespace-stripped exact match
Wall-clock time per question	absolute	seconds	mean after 7200-second display clipping	Not reported
Cost per question	absolute	USD	mean after USD-10 display clipping	Not reported

Results

Model	Metric	Value	n
Claude Code (Sonnet 4.6)	Accuracy	70 percent One full run.	100
Claude Code (Sonnet 4.6)	Wall-clock time per question	1049.2 seconds Printed Figure 2 label.	100
Claude Code (Sonnet 4.6)	Cost per question	1.2 USD Printed Figure 2 label.	100

Evidence

page: PDF pp. 13–14, Methods (Reports Claude Code v2.1.87, claude-sonnet-4-6 1M, high effort, one run, tools, and timeout policy.) — supports /benchmark_version, /scope, /protocol
figure: PDF pp. 3–4, Figure 2A–C (Printed labels report 70.0%, 1049.2 seconds, and USD 1.2.) — supports /metrics, /results

Evaluation run

compbiobench-sonnet-hardest

From Agentic systems are adept at solving well-scoped, verifiable problems in computational biology

Evaluated models / systems: Claude Code (Sonnet 4.6)

Scopesubset · n=17

Shotszero-shot

Turnsmulti-turn agent trajectory

System prompt publicNo

Reasoning / efforthigh

Browserweb access through agent environment

InternetYes

DatabasesYes

Code executionYes

ContainerNo

External toolsYes

Token budgetNot reported

Time / cost budget120 minutes initially; timeout rerun once with 240 minutes

TemperatureNot reported

SeedNot reported

Repeats1

Graderwhitespace-stripped exact string match with occasional manual format-only correction · human review: yes

Statisticsproblem-weighted accuracy in one run over the 17-task subset

Contaminationbenchmark construction reduces lookup shortcuts

Metrics, results, and full protocol

Metrics

Metric	Kind / baseline	Unit	Aggregation	Threshold / tolerance
Accuracy — difficulty Levels 4–5	absolute	percent	problem-weighted mean	whitespace-stripped exact match

Results

Model	Metric	Value	n
Claude Code (Sonnet 4.6)	Accuracy — difficulty Levels 4–5	53 percent Supplementary Figure 2 label; one run.	17

Evidence

page: PDF pp. 13–14, Methods (Exact Claude Code Sonnet configuration and common protocol.) — supports /protocol
figure: PDF p. 11, Supplementary Figure 2B and caption (Levels 4 and 5 grouped; 17 tasks; Sonnet label 53.) — supports /scope, /metrics, /results

/name
/organizations
/kind
/summary
/capabilities
/modalities

compbiobench-zenodo-resource · release: Zenodo immutable record 19443186, published 2026-04-06 (Record title identifies CompBioBench v1 and links the concept DOI.) · Supports 4 fields

Open source →

/release_date
/latest_version
/resources/1/url
/resources/1/pin

compbiobench-zenodo-resource · repository-path: compbiobench.v1.tsv (MD5 b9d72c04c018ee25798cc93ba77c1964) (Contains 100 unique question_id rows; domain counts sum to 100, style counts sum to 100, and internet_required is 78 True / 22 False.) · Supports 8 fields

Open source →

/task_counts/total
/task_counts/basis
/task_counts/subsets
/versions/0/task_counts/total
/versions/0/task_counts/basis
/versions/0/task_counts/subsets
/coverage_notes/0
/scientific_task_classification/entries/0

compbiobench-hf-data-resource · repository-path: compbiobench.v1.tsv at commit 86ee0a22e036fef2a98f382f8fd528d5d390dde3 (Public questions, file paths, domain, style, skills, internet, and GPU flags.) · Supports 4 fields

Open source →

/domains
/coverage_notes/1
/coverage_notes/2
/scientific_task_classification/entries/1

compbiobench-leaderboard-resource · repository-path: app.py and README at commit 6a63e6d2cae531d9e9bf46b341606ed6b304ba3f (Public exact-match submission UI backed by private submissions, ground-truth, and results repositories.) · Supports 6 fields

Open source →

/access/level
/access/grader
/access/license
/resources/4/license
/resources/4/pin
/implementations/1

Agentic systems are adept at solving well-scoped, verifiable problems in computational biology · page: PDF page headers, first-page copyright statement, and Data availability (Preprint is CC BY-NC 4.0 and links the official data, runner, and leaderboard resources.) · Supports 4 fields

Open source →

/access/license
/resources
/resources/0/license
/resources/0/pin

compbiobench-zenodo-resource · release: Zenodo record 19443186 metadata and file manifest (Dataset is CC BY 4.0 and provides the immutable TSV/data-archive files.) · Supports 5 fields

Open source →

/access/tasks
/access/artifacts
/access/license
/resources/1/license
/resources/2/license

compbiobench-runner-resource · repository-path: LICENSE, README.md, run_benchmark.py, and environment.yml at dc350ed37ccd7d7ce96347d139f06dc4bf283f26 (MIT license and per-question Conda execution without grading.) · Supports 5 fields

Open source →

/access/license
/resources/3/license
/resources/3/pin
/implementations/0
/scientific_task_classification/entries/0

View source-level modification history on GitHub →

Access & reproducibility

Access: Partially open
License: CC BY 4.0 for the dataset; MIT for the runner; Apache-2.0 for the leaderboard Space; CC BY-NC 4.0 for the preprint
Grader: The public leaderboard scores whitespace-stripped exact string matches, but ground-truth answ…
Implementations: 2 recorded · official

Access details

Tasks: All 100 v1 questions and their metadata are open on Zenodo and Hugging Face.
Artifacts: The versioned Zenodo record provides a 12,001,320,960-byte input-data archive and a 100-row TSV; a commit-pinned Hugging Face mirror exposes the task files individually.
Grader: The public leaderboard scores whitespace-stripped exact string matches, but ground-truth answers and the evaluation backend are private; the public runner executes agents and extracts answers without grading them.
Biosafety: The registry mirrors no task data or answers. The creators use synthetic/augmented data and scrubbed metadata to prevent lookup shortcuts; source-resource terms still apply.

Resources & implementations

Primary resources

compbiobench-paper-resource · CC BY-NC 4.0
paper · snapshot sha256:29fcc2a5b0b2314a1810b07f338847ca3aa900d7c72c21f98e66d57f19972862 · checked 2026-07-21
compbiobench-zenodo-resource · CC BY 4.0
dataset · version v1-record-19443186 · checked 2026-07-21
compbiobench-hf-data-resource · CC BY 4.0
dataset · commit 86ee0a22e036fef2a98f382f8fd528d5d390dde3 · checked 2026-07-21
compbiobench-runner-resource · MIT
repository · commit dc350ed37ccd7d7ce96347d139f06dc4bf283f26 · checked 2026-07-21
compbiobench-leaderboard-resource · Apache-2.0
leaderboard · commit 6a63e6d2cae531d9e9bf46b341606ed6b304ba3f · checked 2026-07-21

Implementations

CompBioBench runner
official · dc350ed3
CompBioBench leaderboard
official · 6a63e6d2