AI Subscription Usage Limits
Last update: October 28th, 2026
How much of the 5-hour and weekly limit does each model use on a $20 plan?
Same four coding prompts, every model and effort level.
OpenAI: ChatGPT Plus
| Model | Effort | After 4 Prompts 5-hour limit used ($20/mo plan) |
Weekly limit used |
Total time | API price equivalent |
|---|---|---|---|---|---|
| GPT-6.1-Sol | Low |
|
1% | 7m 47s | $0.42 |
| Medium |
|
2% | 10m 40s | $0.45 | |
| High |
|
||||
| GPT-6-Astra | Medium |
|
|||
| High |
|
||||
| GPT-6.1-Terra | Medium |
|
|||
| High |
|
||||
| Xhigh |
|
||||
| Max |
|
||||
| GPT-6-Luna | Medium |
|
|||
| High |
|
||||
| Xhigh |
|
||||
| Max |
|
||||
| Unlock 11 more results with Premium | |||||
Anthropic: Claude Pro
| Model | Effort | After 4 Prompts 5-hour limit used ($20/mo plan) |
Weekly limit used |
Total time | API price equivalent |
|---|---|---|---|---|---|
| Opus 5.5 | Low |
|
1% | 5m 33s | $1.56 |
| Medium |
|
1% | 9m 21s | $2.34 | |
| High |
|
||||
| Sonnet 5.5 | Medium |
|
|||
| High |
|
||||
| Haiku 5.5 | Medium |
|
|||
| High |
|
||||
| Unlock 5 more results with Premium | |||||
This is Premium Content
Subscribe to unlock this full usage table and get access to all premium content.
30-day refund · cancel anytime
Already a member? Sign in
Methodology
- Every model and effort level ran the same four coding prompts. The usage meters in each provider's settings were read before and after the run.
- Providers show usage in whole percents, so <1% means the weekly meter did not move.
- API price equivalent is the cost reported by the coding harness, as if the same tokens were billed via the API.
- Total time depends on the provider's server load at the time of testing, so treat it as a rough indicator.
- Providers change their limits without notice. Check the last update date above.
Looking for quality, not quota? See how these models score on the LLM Coding Leaderboard.
