ModelDial subscription data
Claude Pro vs Max vs Ultra usage limits
Plan pricing, allowance comparisons and usage calculations
Claude
3 plansSwipe to compare →| Weekly usage100M tokens · USD | ProRecommended $20 / mo | Max $100 / mo | UltraRecommended $200 / mo |
|---|---|---|---|
| Sonnet 5.5 | ≈9.69API equiv. ≈ $286.38Per $1 ≈ 210M tokens | ≈46.15API equiv. ≈ $1,365.92Per $1 ≈ 200M tokens | ≈97.62API equiv. ≈ $2,891.31Per $1 ≈ 212M tokens |
| Opus 5.5 | ≈6.69API equiv. ≈ $271.85Per $1 ≈ 145M tokens | ≈32.31API equiv. ≈ $1,321.15Per $1 ≈ 140M tokens | ≈66API equiv. ≈ $2,706Per $1 ≈ 143M tokens |
| Fable 5.2 | No sample | ≈4.6API equiv. ≈ $361.1Per $1 ≈ 20M tokens | ≈9API equiv. ≈ $704.9Per $1 ≈ 20M tokens |
| Fable 5.1 | Not included | ≈3.69API equiv. ≈ $293.77Per $1 ≈ 16M tokens | ≈7.38API equiv. ≈ $573.46Per $1 ≈ 16M tokens |
Usage notes
Average monthly capacity uses weekly allowance × 52 ÷ 12. Monthly plans use their monthly allowance.
OpenAI: Plus and Pro $100/$200/$500 include GPT-6.1 Terra. Plus has a five-hour window; Pro uses weekly limits.
Claude: Pro, Max and Ultra. Max 20× migrates to Ultra. Verified Ultra accounts have no five-hour cap; unverified accounts have a temporary cap. Fable can use up to 50% of the weekly allowance.
GLM: China and international Lite, Pro and Max plans show GLM-5.3 and GLM-5.3-Flash separately for five-hour, weekly and average monthly usage. Both models share the plan allowance.
Kimi: Plus, Pro, Max and Ultra offer K3.1 Pro, K3 and K3-256K. Pro and above offer K3.1 Pro Highspeed. Current plans use five-hour windows and a shared monthly allowance.
Google: Antigravity AI Pro and Ultra 5×/20×. Pro Argon uses a separate weekly allowance; Ultra uses a shared pool.
MiMo: four mainland and four international tiers offer V2.6 Flash and Pro with a shared monthly allowance.
MiniMax: mainland and international Go, Explore and Build tiers show M3.1 Flash capacity for five hours, a week and an average month.
What these plans give you
The same comparison list as By budget. Set conditions once, up to 4 plans.
100% is a full period. For your own plan, enter what is left. Add a monthly usage to see whether each plan is enough. Shared images always compare a full average month.
| Item | ClaudePro $20 / mo |
|---|---|
| Model | |
| Tokens leftWeekly | ≈ 669M |
| API valueWeekly | ≈ $272 |
| Effort | —No effort data |
| Work timeWeekly | —No runtime sample yet |
| Overall scoreout of 100 | 70.3model best |
Work time = API value left ÷ the model’s backend cost per hour; only some models have runtime samples. Without a matching result, the score shows the model’s best published score.
Which Claude tier fits
Only Claude tiers with measured allowances. The green line shows what the next tier adds.
3 tiers · priced in USD
Cheapest first · average month
About these plans
Which Claude model is compared here?
Rows show Sonnet 5.5, Opus 5.5 and Fable 5.1. Fable can use up to 50% of the allowance. Earlier Opus records are available in history.
Where can I find tokens per unit of currency?
Tokens per unit of currency appear below each model’s capacity, calculated as average monthly tokens divided by the paid monthly fee in the selected currency.
Does reaching the five-hour limit exhaust the week?
Pro and Max have separate short and weekly windows. Verified Ultra accounts have no five-hour cap; unverified accounts have a temporary cap. Max 20× automatically migrates to Ultra at the same $200 monthly price.