OpenAI Pro 20×. Implied weekly allowance ≈$1,532 API-list.

Method

Apply published GPT-6 Sol credit-equivalent rates (50 / 5 / 250 credits per million fresh, cached, and output tokens) to the 17-run cohort, then calibrate against a 46.13% seven-day analytics screenshot covering 17,676.36806 credit equivalents. Implied weekly allowance = 2,509.36 / 0.0655 ≈ 38,311 credit equivalents ≈ $1,532 API-list. Labelled implied, not reported. Monthly is that implied weekly figure ×4, a projection.

2,509.36 / 0.0655 = 38,310.83969465649 ≈ 38,311 credit equivalents ≈ $1,532 API-list equivalent. Cross-check: $100.37 / 0.0655 = $1,532.37 ≈ $1,532. Listed conversion: 2,509.36121 × $0.04 = $100.3744484 ≈ $100.37, and 38,311 × $0.04 = $1,532.44 ≈ $1,532. Implied, not a provider-reported quota.

Projection of the implied weekly allowance: ≈$1,532 × 4 = ≈$6,128, and ≈38,311 × 4 = ≈153,244 credit equivalents. Not a separately metered monthly quota, and not four times the cohort.

Lower confidence: the cohort share is a calibrated estimate from a seven-day analytics screenshot and credit weights, not a provider-reported quota debit.

Measured cohort

GPT-6 Sol Max, 17 of 18 runs. 46.13% × 2,509.36121 / 17,676.36806 = 6.5487% of one weekly allowance. The headline share is ~6.55%. The report's planning estimate is ~6.6%. Decimal precision is not measured accuracy.

Fresh input, cached input, and output including reasoning, from the 17 selected GPT-6 Sol results. Rates: 50, 5, and 250 credits per million tokens.

Calibration

Unrounded share

6.5487%

Headline share

~6.55%

Planning estimate

~6.6%

Four identical cohorts

10,037.44484 credit equivalents (~$401.50 API-list), ~26.19 percentage points of one weekly allowance

46.13% × 2,509.36121 / 17,676.36806 = 6.5487% of one weekly allowance

(5,804,476 × 50 + 379,685,632 × 5 + 1,282,837 × 250) / 1,000,000 = 2,509.36121 credit equivalents (~$100.37 at the listed credit conversion)

Supporting observations

Cohort credit equivalents

2,509.36121 (~$100.37 API-list)

Unrounded share

6.5487% of one weekly allowance

Headline share

~6.55%

Planning estimate

~6.6%

Four identical cohorts

~26.19 percentage points of one weekly allowance; 10,037.44484 credit equivalents (~$401.50 API-list)

Published subscription fee

ChatGPT Pro $200 (Pro 20x), separate from this allowance calibration

Caveats

1. These percentages are calibrated estimates, not provider-reported per-benchmark charges or exact quota debits. Token counts are measured; mapping them to a subscription percentage is inferred.
2. The 46.13% Codex Usage Analytics figure is a seven-day analytics view for the long-running task “Find remaining DeepSeek V4 tests”, with zero purchased credits used. That task includes orchestration, subagents, and models other than GPT-6 Sol. 46.13% is not the benchmark cohort's measured consumption. A quota resets on its own weekly boundary; some selected results predate the current quota window.
3. Summing locally available, model-weighted token records for that calibration task and its subagents gives 17,676.36806 credit equivalents. The estimate assumes those records cover exactly the usage attributed to the screenshot, and that included-plan quota is proportional to these credit weights. Local task logs may not cover every item the dashboard attributes to the task.
4. Credit-equivalent rates (50, 5, and 250 credits per million fresh input, cached input, and output tokens) need not match the internal weighting of included Pro usage.
5. Dollar values are API-list equivalents, not subscription cash charges. See OpenAI's pricing and usage explanation at https://learn.chatgpt.com/docs/pricing.
6. Turbofan is not in the 17-run total. There is no final GPT-6 Sol result JSON for it. Space Flight Stage 1 is a prerequisite for Stage 2 and is not a separate row; its independent cost is not added. Retries or preparatory work outside selected result records may also be omitted.
7. Do not publish the decimal precision as measured accuracy. The report's planning estimate for the 17 recorded Sol runs is ~6.6%. The headline calibration stored here is ~6.55% (unrounded 6.5487%).
8. Four identical cohorts are ~26.19 percentage points of one weekly allowance (10,037.44484 credit equivalents, ~$401.50 API-list). Against four equal weekly allowances the fraction remains ~6.55%. The implied weekly allowance of ≈38,311 credit equivalents (≈$1,532) is derived as 2,509.36 / 0.0655 and is not a provider-reported quota.

Evidence

Record status: provisional.

• Source report
• Measurement JSON
• Manifest
• Checksums
• Published RemakeBench results