30

Submitting a run

NEXBENCH3 min read

Mint a manifest, validate it against the shared 12-check intake validator, and submit it to the leaderboard.

A full run against the reference environment pack produces a results draft; nexbench mint turns that draft into a hash-valid manifest by snapping category rates onto the achievable trial grid and computing the run id over the finished results block. You cannot hand-author a valid run id. Minting is the only correct path to one.

  1. 1

    Mint

    Assemble a complete manifest from your run draft (agent identity, submitter, category rates, metrics, trace root).

    $ nexbench mint --from draft.json --out my-run.json
  2. 2

    Validate

    Run all twelve intake checks locally, the exact checks the intake API enforces. Exit code 0 means accepted.

    $ nexbench validate my-run.json
  3. 3

    Submit

    Dry run by default: it validates and prints the destination but sends nothing. Add --yes to actually send.

    $ nexbench submit my-run.json --yes

nexbench submit refuses to send anything that would be rejected, so a submission never leaves your machine unless it would already be accepted.

One validator, three call sites

The twelve-check catalog is one exported list of checks and one validator function, imported verbatim by the submit portal’s browser preview, the intake API, and this documentation, so a rule added there updates all three at once. Error-severity failures block intake outright; warn-severity flags pass but hold the entry for manual review.

#CheckSeverityRejects
1Schema shapeerrorMalformed or mistyped manifests
2Suite & environment pinserrorWrong suite/version/task-count or a modified envPinsDigest
3Metric bounds & budgetserrorRates outside [0,100]; runs over the 900s / $10 per-task budget
4Trial-grid alignmenterrorScores off the achievable m/(N·k) grid (catches fabricated numbers)
5Internal consistencyerrorpass⁵ exceeding weighted pass@1; a wrong category set
6Run-id recomputationerrorAny post-mint edit to a covered field
7Canary attestationerrorContaminated runs: the canary GUID surfaced in output
8Duplicate run / traceerrorResubmitted runIds and replayed trace archives
9Near-duplicate scoreswarnA score vector within one grid step of an existing entry
10Published harness buildwarnUnknown harness images (capped at self-reported)
11Submitter identityerrorMissing a contactable email or HTTPS URL
12Unexpected fieldswarnFields outside the schema, a spoofed-badge vector

Provenance and verification tiers

FieldValuesMeaning
Provenance kindreference · seed · communityRun by Nexis · an illustrative entry · a third-party submission
Verification tierverified · self-reportedNexis re-executed and reproduced the score · the checks passed but were not re-executed

Submission rules

One leaderboard entry per agent + model + scaffold configuration. Ablations are welcome as separate, clearly named entries. Hardcoding task ids, gold answers, or checker logic disqualifies a run outright and does not generalize, since checkers and the private split rotate quarterly. Closed, proprietary agents are fine: point the adapter at an HTTP endpoint and your weights and prompts never leave your infrastructure.

Was this page helpful?

Edit on GitHub