Net Good IndexSubmit a correction

Benefit Ledger · provisional · Economics & Productivity

1Password measures a 21% engineering-productivity gain with Codex

On 8 September 2026 OpenAI published a 1Password case study reporting a 20.9% productivity gain and 10.9% shorter median pull-request cycle time for the core Codex user cohort. 1Password’s own model attributed about 40% of that gain to Codex.

Tier 1 UsefulMethodology 0.1
Submit a correction

Current score

+0.04

1 base · Useful (tier 1 of 5, 1 pts)
× 0.4000 attribution · Material acceleration
× 0.3000 evidence · Technical disclosure or preprint
× 0.8000 realization · Independently validated or deployed
× 0.4500 durability · Medium-term
Event-level product before credit split: 0.04

A measured productivity effect at one firm is useful (tier 1), not an industry-wide launch. Attribution follows 1Password’s own ~40% Codex share. Evidence is a developer case study (0.20). Realization is production use (0.80). Durability is limited because the product and the firm’s workflow will change. Follow-up ingest: independent source raised evidence; added a follow-up source.

What happened

On 8 September 2026 OpenAI published a customer case study stating that 1Password integrated Codex across planning, implementation, review, testing, and production investigation. For its core Codex user cohort 1Password reported a 20.9% engineering-productivity improvement and a 10.9% reduction in median pull-request cycle time. One cited example is a multi-service defect investigation falling from about two hours to 5–20 minutes. The figures are vendor-reported at a single company. 1Password’s own ROI model assigned only about 40% of the measured productivity gain to Codex. This record is a measured workplace outcome, scored separately from the 2025 Codex product launch. Update (2026-09-17): Finding Cloud Misconfigurations with Frontier AI: A Case Study. Finding Cloud Misconfigurations with Frontier AI: A Case Study

Model attribution

+0.04

Codex

Software-engineering agent used across 1Password’s delivery lifecycle.

The case study names Codex, OpenAI’s software-engineering agent. Mapped to the named codex-1 release in that family until a later snapshot is identified.

Attribution 0.4000 · Credit share 100% · OpenAI

Claims

  • 1Password measures a 21% engineering-productivity gain with Codex

    outcome · supported

Sources

primary sources

independent sources

Revision history

  • 13 Sep 2026 · 0.00 → 0.00

    Auto-published from news ingest.

  • 14 Sep 2026 · 0.00 → 0.03

    Editorial review of a news-ingest placeholder: rewritten, attributed to a named system, and rescored.

  • 17 Sep 2026 · 0.03 → 0.04

    Ingest update: independent source raised evidence; added a follow-up source.

Submit a correction