Skip to content

Compare

Two results on one task. The same model slug means only the harness varies.

taskboardtask 1.0.0 · ruleset 1.0.0 · scoring 1.0.09 results on this task

A attribution

run passed the threshold; no failure to attribute

no proof

B attribution

none

no proof

Run metric deltas, B minus A
MetricABB − A
Steps160-16
Tool calls160-16
Tool error rate000
Total tokens92,5250-92,525
Cost USD0.0640-0.064
Duration ms66,1050-66,105
Different modelThe harness and the model both differ, so the score gap is not attributable to one.

A

Organizer-run

OpenAI Codex CLI (openai/gpt-5.1-codex)

openai/gpt-5.1-codex

100.0

Spec
100.0
Security
100.0
Maintainability
81.7
4 pass0 fail0 unverified
Open result

B

Organizer-run

Aider 0.86.2 (openai/gpt-4o-mini)

openai/gpt-4o-mini

97.3

Spec
100.0
Security
100.0
Maintainability
82.0
9 pass0 fail0 unverified
Open result
Bench deltas, B minus A
BenchABB − A
Overall100.097.3-2.7
Spec100.0100.00.0
Security100.0100.00.0
Maintainability81.782.0+0.3
Requirement outcomes for both results
RequirementBenchAB
REQ-AUTH-001Project access must be owner-scopedspecUNVERIFIEDPASS
REQ-DEP-001No shell execution dependencysecurityUNVERIFIEDPASS
REQ-ERR-001Project API returns structured error responsesspecUNVERIFIEDPASS
REQ-PATH-001Projects API route existsspecUNVERIFIEDPASS
REQ-SEC-002No environment file in the repositorysecurityUNVERIFIEDPASS
REQ-SEC-003No dynamic code evaluation in the project routesecurityUNVERIFIEDPASS
REQ-TASKS-001Task list API route existsspecUNVERIFIEDPASS
REQ-TASKS-002Task list reads through the repository layerspecUNVERIFIEDPASS
REQ-UI-001Task list component existsspecUNVERIFIEDPASS