s42chen/sci-run

Science Speedrun: 21 completed Astra task histories

Reproducible submitted solutions, native evaluation reports, and screened Codex logs from GPT-6 Astra with xhigh reasoning. These are file artifacts, not a tabular training dataset. The histories contain 540 inference-distinct candidates; each has a completed sealed evaluation. Existing additional measurements of duplicate packages are retained. reviews/fairness-20261031/ holds the operator-side fairness review of every first-round trajectory (per-task verdicts and the cross-cutting infrastructure findings); each first-round index carries its verdict under operator_fairness_review, and a fresh trial's index names the first trial it repeats under fresh_trial_of.

Task	Distinct candidates	Artifacts
atomistic-materials	3	Files
brain-computer-interface-decoding	127	Files
crystal-structure-generation	20	Files
earth-observation-pretraining-transfer	11	Files
fluid-flow-control	11	Files
fruit-fly-neuron-reconstruction	7	Files
genome-sequence-to-function	33	Files
mass-spectrometry-molecule-retrieval	21	Files
molecular-pretraining-transfer	14	Files
natural-disease-progression	13	Files
pde-surrogate	60	Files
probabilistic-global-weather	9	Files
protein-design	1	Files · superseded by the corrected trial
protein-design	3	Files
protein-folding-prediction	18	Files
protein-sequence-modeling	58	Files
protein-sequence-modeling	9	Files · fresh trial after the fairness review
single-cell-perturbation-prediction	17	Files
surface-code-decoding	45	Files
timeseries-forecasting	53	Files
tokamak-magnetic-control	7	Files

Layout

tasks/<task>/<run-id>/ contains submissions/, evaluations/, logs/, finalization/, and one index.json. The index maps every original snapshot and research timestamp to its canonical package and actual evaluation, and records file hashes, executable permissions, source commit, images, resources, and deviations. plots/ contains the complete-history figures and their metric/reference CSV files. reproduction/files/ preserves the frozen verifier/adapter source files referenced by each evaluation. Scientific native reports are retained without rewriting scores. Fruit Fly evaluations also keep the final SHAPE stage's per-node predictions (evaluations/<snapshot>/disjoint-pair-shape/node-assignments.npz: skeleton node_id to predicted segment prediction_id, with that stage's metrics.json), so other segmentation metrics can be computed without rerunning inference.

Disease submissions include their required Python/CUDA runtime. Identical large weights and bundled dependencies are shared under shared/package-files/, as plain files, so repeated checkpoint contents need only be downloaded once. Small entry-point files remain browsable in each submission folder. Indexes map stored files back to their original paths and executable permissions. Reconstruct a package before execution. From a complete downloaded repository:

python restore_submission.py tasks/<task>/<run-id>/index.json S001 /path/to/new/submission

To fetch just one candidate, download its index.json and the helper first (keeping the directory layout), then add --repo s42chen/sci-run to the command above. With huggingface_hub installed, the helper downloads only that candidate's missing files. Use --revision <dataset-commit> to pin the download to a specific dataset version.

Use GRACE_FINAL for the separate post-budget package. The helper copies files into a new directory, restores permissions, and checks the exported names/modes/content fingerprint. It never executes the submitted code. Do not execute submissions merely to inspect this archive; use the benchmark's isolated evaluation environment.

Interpretation and limitations

All runs used one accessible H100 and 48 hours of charged research time. All tasks except Tokamak, Disease, Earth, and BCI were run with reduced CPU/RAM/GPU allocations relative to their pinned eight-GPU task configurations; those four match their tasks' one-GPU research requests at their recorded source commits. Runs were interrupted and resumed (operator pauses, Beaker preemptions, one shared-storage outage); preserving the clock does not reproduce uninterrupted computation. The fairness review under reviews/ quantifies the operator-caused losses per task and names the trials judged materially unfair. Fresh trials of those tasks were run from scratch (new session, same images, data, and budget) and are added as separate run directories next to the retained first trials, as each fresh history is fully scored: Protein Sequence (run 20261031a) is included; Weather, Genome, Atomistic Materials, and Mass Spectrometry will follow. Brain-computer interface decoding (BCI), the nineteenth task, ran once at its native budget; its five tracks are scored and plotted separately. Beaker execution uses documented isolation/resource adaptations and may grant more RAM than requested. Published baseline lines are reference measurements, not equal- budget human trials. Missing or incompatible baselines are explicitly omitted.

The first Protein Design trial (run 20261022a) ran without its pinned pretrained inference assets mounted; it is retained for the record but superseded by the corrected trial (run 20261027a), which is the Design result shown in the plots. finalization/ is an additional submission-only window of at most five minutes; it is not backdated into the 48-hour curves. Purple markers denote those reviewed post-budget results; orange marks denote reused scores for duplicate snapshots.

Two results carry operator rulings (2026-11-02). Tokamak is scored under the task's metric v3 that the trial was pinned to (35 sealed episodes, timed-out episodes voided, so per- candidate denominators vary from 32 to 35); the task specification moved to v4 (105 episodes, no voiding) after the trial, and no rescoring was performed. The PDE surrogate final package, from its sixteenth candidate onward, routes most datasets by case metadata to independently implemented equation operators declared in its manifest; this is permitted by the task contract and the scores stand, but from that point the reward mostly measures re-derivation of the governing equations rather than a learned surrogate.

Provenance and excluded material

Submissions were produced by the evaluated agent; reports by the native scientific scorers under the recorded adapters; logs by Codex and the harness. Organization, indexes, plots, and the reconstruction helper were created by the operator. Logs are screened for credentials, with redaction counts and original/exported hashes recorded in the indexes. Authentication stores, private test inputs/labels, raw research workspaces and unsubmitted checkpoints are excluded. Dataset inputs and separately distributed pretrained dependencies remain at their upstream locations. Existing source and dependency licenses continue to apply; no blanket relicensing is asserted here.

Benchmark source: https://github.com/sunnweiwei/science-speedrun

Lean export

Mass Spectrometry packages omit unused research backups and unselected generations. The selection follows the same reviewed inference footprint used to deduplicate evaluations. Every retained inference file, mode, and relevant verifier metadata was checked against the recorded evaluated identity; required notes and adjacent licenses are also retained. No model code, weights, paths, or scores were rewritten. These trimmed packages have their own export fingerprint, with the original full-package fingerprint and evaluated inference identity recorded separately in each index. Other tasks retain their exact submitted package contents, with identical files shared. Original full run artifacts remain preserved on the operator's storage.