rewirebio.iobenchmarks
Evaluation

Kimi K3 on BenchBench-Cloning construct design: Minimal-edit recoding

Reported measurement transcribed from the pinned preprint. Not independently reproduced; the benchmark code and task data were not retrievable.

Evaluation results

1 evaluation · 1 result. Different protocols are not a single leaderboard.

Filter evaluations

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configurationProtocol and datasetFindingEvidence and details
Configuration: Kimi K3 in Pi 1.0.3, high reasoning effortTask: BenchBench-Cloning construct design: Minimal-edit recoding
Dataset subset: BenchBench-Cloning construct-design tasks with a rubric model judge (10)
5/5 Count (construct-design attempts passed out of 5 on one task at top reasoning effort, pass at a final score of at least 0.9; grading deterministic multiplied by a GPT-6 Astra rubric score, so the outcome depends on the judge; top reasoning tier, the highest of this system's three settings)
count · higher

Uncertainty: Not reported by the source: Table G2 prints no interval for the per-task counts.

Coverage: 5/5 attempts scored

Author-reported evaluation · Source checked
Methods, coverage and source

Kimi K3 on BenchBench-Cloning construct design: Minimal-edit recoding

benchbench-cloning-20261106-protocol-construct-design-per-task-rubric-judged

Aggregation: Not reported

BenchBench-Cloning: Evaluating Molecular Cloning Design and Protocol Development (arXiv:2610.33355v1) · Table G2, row 'Minimal-edit recoding', column 'Kimi K3'

Source checking is not independent reproduction. Release 2026-11-21-c8bea25e5abb.

Research readiness

0 of 4 readiness checks met. These checks assess whether the evidence supports a reproducible investigation; a source-checked score alone does not meet them.

Readiness checks, gaps and artifacts

Release 2026-11-21-c8bea25e5abb · Evidence verified: Not verified

Evidence incomplete

Replay metrics

Exact outcomes, predictions, identifiers and evaluator are connected.

Missing or unresolved evidence

  • No verified artifact manifest is linked to this exact record.
  • File checksums match the recorded files: not yet verified
  • Predictions are matched to the right samples: not yet verified
  • Score meaning and direction are confirmed: not yet verified
  • Metrics are recomputed from the saved predictions: not yet verified

Verified: Not verified

Evidence incomplete

Investigate discrepancies

Replay evidence includes annotations and an assessment of dependence. Unknown independence permits descriptive analysis only.

Missing or unresolved evidence

  • No verified artifact manifest is linked to this exact record.
  • File checksums match the recorded files: not yet verified
  • Predictions are matched to the right samples: not yet verified
  • Score meaning and direction are confirmed: not yet verified
  • Metrics are recomputed from the saved predictions: not yet verified
  • Sample annotations are recorded: not yet verified
  • Dependence between samples is assessed: not yet verified

Verified: Not verified

Evidence incomplete

Run locally

A pinned recipe describes the inputs, environment and resource requirements.

Missing or unresolved evidence

  • No verified artifact manifest is linked to this exact record.
  • File checksums match the recorded files: not yet verified
  • Predictions are matched to the right samples: not yet verified
  • Score meaning and direction are confirmed: not yet verified
  • A pinned run recipe exists: not yet verified
  • Compute requirements are estimated: not yet verified

Verified: Not verified

Evidence incomplete

Validate independently

Separate data and exposure records support an independent test.

Missing or unresolved evidence

  • No verified artifact manifest is linked to this exact record.
  • File checksums match the recorded files: not yet verified
  • Predictions are matched to the right samples: not yet verified
  • Score meaning and direction are confirmed: not yet verified
  • Independent validation data exist: not yet verified
  • Overlap with training data is checked: not yet verified

Verified: Not verified

Readiness describes the evidence in this release. Availability on your computer is checked separately when an investigation runs. Existing data exposure can prevent independent validation even when files are available.

Artifacts and reproduction

No verified artifact manifest is connected to this record yet. The gaps above identify what is needed before analysis can begin.

Evaluation procedure

Configuration
Kimi K3 in Pi 1.0.3, high reasoning effort
Task
BenchBench-Cloning construct design: Minimal-edit recoding
Dataset subset
BenchBench-Cloning construct-design tasks with a rubric model judge (10)
origin
Author-reported evaluation
configuration
arXiv:2610.33355v1 as retrieved 2026-11-06
adaptation
Single-turn agent attempt: the agent inspects the supplied files, uses the available tools, revises within the attempt and writes its submission to a designated answer file; no clarifying questions and no feedback between attempts
aggregation
None: the count is per task and per system
budget
Modal sandbox, 2 vCPU and 6 GiB memory, network restricted to the model provider, two hours per attempt
dataset version
BenchBench-Cloning task corpus as published in arXiv:2610.33355v1
inputs
Scientist's objective, sequence records and supporting materials; for six tasks also experimental records from the source study
metric implementation
Final score of at least 0.9, the deterministic grader's result multiplied by a GPT-6 Astra rubric score of the written reasoning and plan
population
One construct-design task ('Minimal-edit recoding'), five attempts
protocol id
benchbench-cloning-20261106-protocol-construct-design-per-task-rubric-judged
split
Evaluation only; this one task scored

Metadata review: source checked. Unreported conditions prevent automatic comparisons.

Reproduction

Split
Evaluation only; this one task scored
Adaptation
Single-turn agent attempt: the agent inspects the supplied files, uses the available tools, revises within the attempt and writes its submission to a designated answer file; no clarifying questions and no feedback between attempts
Scoring implementation
Final score of at least 0.9, the deterministic grader's result multiplied by a GPT-6 Astra rubric score of the written reasoning and plan

No execution recipe has been verified for this exact configuration and evaluation. A benchmark's general instructions may use different inputs, splits or model settings.

Reproducing this published result requires matching its model configuration, data, split and scorer. Source checking or a successful smoke test does not establish score reproduction.

Evidence

Source checking verifies the cited claim or transcription. It does not establish independent reproduction.

Evidence table

Inspect claims, sources and review details

Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.

One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.

24 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-11-21-c8bea25e5abb
Property and statementOriginal source and locationReview and provenance
Comparison: adaptation
Single-turn agent attempt: the agent inspects the supplied files, uses the available tools, revises within the attempt and writes its submission to a designated answer file; no clarifying questions and no feedback between attempts
Context-only references
BenchBench-Cloning: Evaluating Molecular Cloning Design and Protocol Development (arXiv:2610.33355v1)

Original source ↗

Table G2, row 'Minimal-edit recoding', column 'Kimi K3'

Version: arXiv v1, dated 23 October 2026
Retrieved: 2026-11-06T11:51:23Z

not individually reviewed

No individual claim review recorded

author reported

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.comparison.adaptation

Source artifact SHA-256: 6f5044b1bbbd81ad86b23cb536e2cae6428e0d9126f11546041355518febdc39

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/arxiv-2610.33355v1.layout.txt.gz

Comparison: aggregation
None: the count is per task and per system
Context-only references
BenchBench-Cloning: Evaluating Molecular Cloning Design and Protocol Development (arXiv:2610.33355v1)

Original source ↗

Table G2, row 'Minimal-edit recoding', column 'Kimi K3'

Version: arXiv v1, dated 23 October 2026
Retrieved: 2026-11-06T11:51:23Z

not individually reviewed

No individual claim review recorded

author reported

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.comparison.aggregation

Source artifact SHA-256: 6f5044b1bbbd81ad86b23cb536e2cae6428e0d9126f11546041355518febdc39

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/arxiv-2610.33355v1.layout.txt.gz

Comparison: budget
Modal sandbox, 2 vCPU and 6 GiB memory, network restricted to the model provider, two hours per attempt
Context-only references
BenchBench-Cloning: Evaluating Molecular Cloning Design and Protocol Development (arXiv:2610.33355v1)

Original source ↗

Table G2, row 'Minimal-edit recoding', column 'Kimi K3'

Version: arXiv v1, dated 23 October 2026
Retrieved: 2026-11-06T11:51:23Z

not individually reviewed

No individual claim review recorded

author reported

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.comparison.budget

Source artifact SHA-256: 6f5044b1bbbd81ad86b23cb536e2cae6428e0d9126f11546041355518febdc39

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/arxiv-2610.33355v1.layout.txt.gz

Comparison: dataset version
BenchBench-Cloning task corpus as published in arXiv:2610.33355v1
Context-only references
BenchBench-Cloning: Evaluating Molecular Cloning Design and Protocol Development (arXiv:2610.33355v1)

Original source ↗

Table G2, row 'Minimal-edit recoding', column 'Kimi K3'

Version: arXiv v1, dated 23 October 2026
Retrieved: 2026-11-06T11:51:23Z

not individually reviewed

No individual claim review recorded

author reported

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.comparison.dataset_version

Source artifact SHA-256: 6f5044b1bbbd81ad86b23cb536e2cae6428e0d9126f11546041355518febdc39

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/arxiv-2610.33355v1.layout.txt.gz

Comparison: inputs
Scientist's objective, sequence records and supporting materials; for six tasks also experimental records from the source study
Context-only references
BenchBench-Cloning: Evaluating Molecular Cloning Design and Protocol Development (arXiv:2610.33355v1)

Original source ↗

Table G2, row 'Minimal-edit recoding', column 'Kimi K3'

Version: arXiv v1, dated 23 October 2026
Retrieved: 2026-11-06T11:51:23Z

not individually reviewed

No individual claim review recorded

author reported

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.comparison.inputs

Source artifact SHA-256: 6f5044b1bbbd81ad86b23cb536e2cae6428e0d9126f11546041355518febdc39

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/arxiv-2610.33355v1.layout.txt.gz

Comparison: metric implementation
Final score of at least 0.9, the deterministic grader's result multiplied by a GPT-6 Astra rubric score of the written reasoning and plan
Context-only references
BenchBench-Cloning: Evaluating Molecular Cloning Design and Protocol Development (arXiv:2610.33355v1)

Original source ↗

Table G2, row 'Minimal-edit recoding', column 'Kimi K3'

Version: arXiv v1, dated 23 October 2026
Retrieved: 2026-11-06T11:51:23Z

not individually reviewed

No individual claim review recorded

author reported

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.comparison.metric_implementation

Source artifact SHA-256: 6f5044b1bbbd81ad86b23cb536e2cae6428e0d9126f11546041355518febdc39

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/arxiv-2610.33355v1.layout.txt.gz

Comparison: population
One construct-design task ('Minimal-edit recoding'), five attempts
Context-only references
BenchBench-Cloning: Evaluating Molecular Cloning Design and Protocol Development (arXiv:2610.33355v1)

Original source ↗

Table G2, row 'Minimal-edit recoding', column 'Kimi K3'

Version: arXiv v1, dated 23 October 2026
Retrieved: 2026-11-06T11:51:23Z

not individually reviewed

No individual claim review recorded

author reported

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.comparison.population

Source artifact SHA-256: 6f5044b1bbbd81ad86b23cb536e2cae6428e0d9126f11546041355518febdc39

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/arxiv-2610.33355v1.layout.txt.gz

Comparison: protocol id
benchbench-cloning-20261106-protocol-construct-design-per-task-rubric-judged
Context-only references
BenchBench-Cloning: Evaluating Molecular Cloning Design and Protocol Development (arXiv:2610.33355v1)

Original source ↗

Table G2, row 'Minimal-edit recoding', column 'Kimi K3'

Version: arXiv v1, dated 23 October 2026
Retrieved: 2026-11-06T11:51:23Z

not individually reviewed

No individual claim review recorded

author reported

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.comparison.protocol_id

Source artifact SHA-256: 6f5044b1bbbd81ad86b23cb536e2cae6428e0d9126f11546041355518febdc39

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/arxiv-2610.33355v1.layout.txt.gz

Comparison: split
Evaluation only; this one task scored
Context-only references
BenchBench-Cloning: Evaluating Molecular Cloning Design and Protocol Development (arXiv:2610.33355v1)

Original source ↗

Table G2, row 'Minimal-edit recoding', column 'Kimi K3'

Version: arXiv v1, dated 23 October 2026
Retrieved: 2026-11-06T11:51:23Z

not individually reviewed

No individual claim review recorded

author reported

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.comparison.split

Source artifact SHA-256: 6f5044b1bbbd81ad86b23cb536e2cae6428e0d9126f11546041355518febdc39

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/arxiv-2610.33355v1.layout.txt.gz

Coverage: eligible
5
Context-only references
BenchBench-Cloning: Evaluating Molecular Cloning Design and Protocol Development (arXiv:2610.33355v1)

Original source ↗

Table G2, row 'Minimal-edit recoding', column 'Kimi K3'

Version: arXiv v1, dated 23 October 2026
Retrieved: 2026-11-06T11:51:23Z

not individually reviewed

No individual claim review recorded

author reported

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.coverage.eligible

Source artifact SHA-256: 6f5044b1bbbd81ad86b23cb536e2cae6428e0d9126f11546041355518febdc39

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/arxiv-2610.33355v1.layout.txt.gz

How to read the evidence

Sources and history

Release 2026-11-21-c8bea25e5abb · Record review: source checked

1 source record and release historyDownload this release (gzip)
Technical metadata and extraction receipts

Stable ID: benchbench-cloning-20261106-eval-kimi-k3-task-minimal-edit-recoding

areas
dna-genomes
contexts
research
comparison
adaptation: Single-turn agent attempt: the agent inspects the supplied files, uses the available tools, revises within the attempt and writes its submission to a designated answer file; no clarifying questions and no feedback between attempts; aggregation: None: the count is per task and per system; budget: Modal sandbox, 2 vCPU and 6 GiB memory, network restricted to the model provider, two hours per attempt; dataset version: BenchBench-Cloning task corpus as published in arXiv:2610.33355v1; inputs: Scientist's objective, sequence records and supporting materials; for six tasks also experimental records from the source study; metric implementation: Final score of at least 0.9, the deterministic grader's result multiplied by a GPT-6 Astra rubric score of the written reasoning and plan; population: One construct-design task ('Minimal-edit recoding'), five attempts; protocol id: benchbench-cloning-20261106-protocol-construct-design-per-task-rubric-judged; split: Evaluation only; this one task scored
coverage
eligible: 5; scored: 5; unit: attempts
denominator
5
limitations
Graded in part by GPT-6 Astra as a rubric judge, which is also one of the eight evaluated systems.; Top reasoning setting only; the paper prints no per-task counts for the lower settings.; Reported by the benchmark's authors, who also built the benchmark and ran every system on it.; Not reproduced by Rewire: The construct-design tasks, graders and reference solutions are said in footnote 1 (page 3) to be available at https://github.com/benchling/benchbench-cloning. That URL returned HTTP 404 from both the web page and the GitHub API on 2026-11-06 at 2026-11-06T11:51:23Z, a repository search for 'benchbench-cloning' returned no results, and the repository is absent from benchling's public repositories. No code or data artifact was retrieved.
n runs
5
origin
author_reported
protocol
benchbench-cloning-20261106-protocol-construct-design-per-task-rubric-judged
review
artifact sha256: 6f5044b1bbbd81ad86b23cb536e2cae6428e0d9126f11546041355518febdc39; date: 2026-11-06; method: deterministic-table-parse; pdf-text-parse; source-hash-verification; method note: Extraction: parsed from the pdftotext -layout text layer of the pinned PDF by extract/parse_tables.py, which locates each table by its own caption line and searches only the lines between the previous caption and that caption, because Figures 2, 3, 6 and E1 print the same numbers as Tables 2, G2 and E1. Every row label is asserted. extract/check_tables.py reparses the same cells independently, by column position rather than by row regex, and compares.; note: Transcribed from the pinned preprint by extract/parse_tables.py; not rerun.; retrieval url: https://arxiv.org/pdf/2610.33355v1; reviewer: claude; reviewer note: Claude (Opus 5) extraction agent. Extraction only: no independent review is claimed, and a different agent reviews this batch.
source locator
Table G2, row 'Minimal-edit recoding', column 'Kimi K3'
version
arXiv:2610.33355v1 as retrieved 2026-11-06
Related records

Suggest a correction