Plan guide

How much coding do you get from Codex Plus?

Available models: GPT-6 Astra / GPT-6.1 Sol / GPT-6 Luna. $20/mo buys about 4 h/week of building at p50 peak (p10–p90 2–8) (source: community). Coding-agent quality scores (/100): GPT-6.1 Sol (high) · Codex · 60/100 planning, GPT-6.1 Sol (high) · Codex · 60/100 building, GPT-6.1 Sol (high) · Codex · 60/100 admin. Building model cost: $0.89 per benchmark task at API prices. Catalog note: Astra p50 ≈ 1h. Research treated Codex messages as model calls — too harsh; these are mid.

Hours by role

p10 / p50 / p90, h/week, at the plan's own clock. Peak and off-peak differ only where the vendor publishes a night rate.

RoleBenchmark quality /100Peak p10 / p50 / p90 (h/week)Off-peak p10 / p50 / p90 (h/week)
PlanningGPT-6.1 Sol (high) · Codex · 60/1001 / 1 / 2—
BuildingGPT-6.1 Sol (high) · Codex · 60/1002 / 4 / 8—
AdminGPT-6.1 Sol (high) · Codex · 60/1003 / 5 / 10—

Clock: no published off-peak discount. Runs as a CLI agent. Source: community.

Get Codex Plus →Rank it against your stack →

Current coding-agent evaluations

The ranking uses the fixed setting named in each role above. These scores do not establish subscription hours. Model availability verified 2026-11-09: vendor documentation.