LanguageEN简中

Subscription Usage Benchmarks

Compare plan capacity and API-equivalent value for the models you use.

21 plans · Updated 2026-10-08View model scores

Models share the plan allowance; rows are not additive. Tokens include cache. Values are usage estimates.

OpenAI

5 plans

Only Pro $200 has new and previous quotas. Plus and Pro $100 are unchanged; Pro $500 is a new tier. The source reports pre-cut Pro $200 plans retain old limits until Oct 29.

Pro 200 · new quota

Monthly fee$200
ModelWeekly100M tokensWeekly API equivalentUSD
GPT-6.1 Sol≈ 23.54≈ $481
GPT-6 Astra≈ 4.38≈ $669

Pro 20x · previous quota

Monthly fee$200
ModelWeekly100M tokensWeekly API equivalentUSD
GPT-6 Astra≈ 8.5≈ $1,313
GPT-6.1 Sol≈ 47.83≈ $942
GPT-6 Sol≈ 52.19≈ $1,329
GPT-5.6 Sol≈ 35.9≈ $1,880

Pro 500

Monthly fee$500
ModelWeekly100M tokensWeekly API equivalentUSD
GPT-6.1 Sol≈ 60.69≈ $1,243
GPT-6 Astra≈ 10.62≈ $1,605

Claude

6 plans

Current estimates use the study’s agentic workload; Fable is limited to 50% of the allowance. Historical samples remain for comparison, not addition.

Pro · new quota

Monthly fee$20
ModelWeekly100M tokensWeekly API equivalentUSD
Sonnet 5.5≈ 9.69≈ $286
Opus 5.5≈ 6.69≈ $272

Max 5x · new quota

Monthly fee$100
ModelWeekly100M tokensWeekly API equivalentUSD
Sonnet 5.5≈ 46.15≈ $1,366
Opus 5.5≈ 32.31≈ $1,321
Fable 5.1≈ 3.69≈ $294

Max 20x · new quota

Monthly fee$200
ModelWeekly100M tokensWeekly API equivalentUSD
Sonnet 5.5≈ 97.62≈ $2,891
Opus 5.5≈ 66≈ $2,706
Fable 5.1≈ 7.38≈ $573

Pro · previous quota

Monthly fee$20
ModelWeekly100M tokensWeekly API equivalentUSD
Opus 5.5≈ 7≈ $230

Max 5x · previous quota

Monthly fee$100
ModelWeekly100M tokensWeekly API equivalentUSD
Opus 5.5≈ 41.62≈ $1,456

Max 20x · previous quota

Monthly fee$200
ModelWeekly100M tokensWeekly API equivalentUSD
Opus 5.5≈ 58.82≈ $2,237

Grok

3 plans

SuperGrok

Monthly fee$30
ModelWeekly100M tokensWeekly API equivalentUSD
Grok 4.7≈ 1.12≈ $45

SuperGrok Plus

Monthly fee$100
ModelWeekly100M tokensWeekly API equivalentUSD
Grok 4.7≈ 2.94≈ $121

SuperGrok Heavy

Monthly fee$300
ModelWeekly100M tokensWeekly API equivalentUSD
Grok 4.7≈ 9.17≈ $401

GLM

5 plans

Current estimates use the official 95% cache-hit range, all peak to all off-peak. Months use 52/12, excluding promotions ending Oct 7. API equivalents remain unavailable.

Lite · new quota

¥118 / month
Monthly fee—
ModelWeekly100M tokensWeekly API equivalentUSD
GLM-5.3≈ 0.48–0.97—
GLM-5.3-Flash≈ 1.46–2.92—

Pro · new quota

¥538 / month
Monthly fee$80
ModelWeekly100M tokensWeekly API equivalentUSD
GLM-5.3≈ 2.9–5.8—
GLM-5.3-Flash≈ 8.77–17.55—

Max · new quota

¥1078 / month
Monthly fee$160
ModelWeekly100M tokensWeekly API equivalentUSD
GLM-5.3≈ 6.76–13.52—
GLM-5.3-Flash≈ 20.47–40.95—

Pro · previous quota

¥538 / month
Monthly fee$80
ModelWeekly100M tokensWeekly API equivalentUSD
GLM-5.3 · V3≈ 9.95≈ $259

Max · previous quota

¥1078 / month
Monthly fee$160
ModelWeekly100M tokensWeekly API equivalentUSD
GLM-5.3 · V3≈ 23.2≈ $603

Google

1 plans

AI Pro

Monthly fee$19.99
ModelWeekly100M tokensWeekly API equivalentUSD
Mixed models≈ 14.36≈ $214

Kimi

1 plans

Allegretto

¥199 / month
Monthly fee$28.03
ModelWeekly100M tokensWeekly API equivalentUSD
K3≈ 3.4≈ $137

Quota calculator

Calculating forOpenAI · Plus

GPT-6 Astra · max · Weekly

Reasoning effort
Period
%

Plus · GPT-6 Astra · Weekly

Available work time
≈ 10.5hours

max · Cumulative model runtime

Overall score95.5out of 100
Remaining tokens≈ 0.43 × 100M tokens
API equivalent≈ $66.00

Tradeoffs within this plan

GPT-6.1 Solmax
Score 93.8 -1.7 pts≈ 32.3 h +21.9 h（+209%）
Compare

CAPACITY IN CONTEXT

How far does the same budget go?

Average-month capacity. Models share each plan’s allowance and are not additive. Charts use the monthly budget below and share one scale.

Capacity at the same price · $200

  • Sonnet 5.5 · Claude$12,529
  • Opus 5.5 · Claude$11,726
  • Fable 5.1 · Claude$2,485
  • GPT-6.1 Sol · OpenAI$2,084
  • GPT-6 Astra · OpenAI$2,897

Model allowances are shown separately and are not additive. Fable uses up to 50% of the allowance.

OpenAI $200: new and previous quotas

  • GPT-6 Astra · New quota$2,897
  • GPT-6 Astra · Previous quota$5,691
  • GPT-6.1 Sol · New quota$2,084
  • GPT-6.1 Sol · Previous quota$4,083

Historical values are preserved. Sample periods and workloads differ, so no percentage change is inferred.

GLM: peak and off-peak

95% cache-hit rate · average month · 100M tokens. Endpoints represent all-peak and all-off-peak usage, not statistical uncertainty.

  • Lite · GLM-5.32.08–4.2
  • Lite · GLM-5.3-Flash6.33–12.65
  • Pro · GLM-5.312.57–25.13
  • Pro · GLM-5.3-Flash38–76.05
  • Max · GLM-5.329.29–58.59
  • Max · GLM-5.3-Flash88.7–177.45