subsdaddy
The four LLM subscriptions I pay for: what each one actually runs, what it costs per billion tokens, and how fast it is.
quota read 2026-10-30 18:27Z · traffic 7 days (2026-10-23 18:39Z to 2026-10-30 18:27Z) · fees owner-confirmed
| Lane | Fee/mo | Quota window | $/B | Tokens 7d | tok/s | TTFT p50 |
|---|---|---|---|---|---|---|
| CommandCode - GOAT | $10.00/mo | monthly 50.0% | $2.61measured | 763M | 214measured | 2,778 msmeasured |
| OpenCode Go | $10.00/mo | monthly 90.5% | $4.24measured | 1.79B | 161measured | 2,407 msmeasured |
| Z.ai Coding Plan - Max (Legacy v2) | $56.00/mo | weekly 67.0% | $14.20measured | 664M | 76measured | 5,064 msmeasured |
| QwenCloud - Token Plan Individual | $10.00/mo | monthly 57.9% | $22.12measured | 475M | 148measured | 3,867 msmeasured |
$/B is the plan fee over the tokens its binding quota window delivered, priced from my gateway's own logs. A lane with no readable window keeps its row and says why.
CommandCode - GOAT
$10.00/mo5h 0.0% · weekly 0.1% · monthly 50.0% · resets 2026-11-21 15:49Z
- Capacity
- 3.83B per 30d · monthly window · published $80.00
- $/B
- $2.61
- Traffic, last 7 days
- 763M tokens · 3,099 requests · cache 97.7% · 7 days
- Speed
- 214 tok/s p50 · 2,778 ms TTFT p50 · 8,524 ms p90
| Model | Share | Tokens 7d | Requests | Cache | TTFT p50 | tok/s | Monthly limit |
|---|---|---|---|---|---|---|---|
| DeepSeek V4.1 Flash | 83.2% | 635M | 2,470 | 97.9% | 2,940 ms | 226 | N/Ano per-model limit |
| DeepSeek V4.1 Flash (fast) | 10.4% | 79.0M | 441 | 98.3% | 1,261 ms | 186 | N/Ano per-model limit |
| MiMo v2.6 Flash | 6.4% | 48.6M | 185 | 93.4% | 5,765 ms | 43 | N/Ano per-model limit |
| Other models (1) | <0.1% | 261K | 3 | 43.6% | N/Apooled | N/Apooled | N/Apooled, under 1% each |
OpenCode Go
$10.00/mosingle_account_reading
5h 9.6% · weekly 81.0% · monthly 90.5% · resets 2026-11-19 00:27Z
- Capacity
- 2.36B per 30d · monthly window
- $/B
- $4.24
- Traffic, last 7 days
- 1.79B tokens · 10,632 requests · cache 88.7% · 7 days
- Speed
- 161 tok/s p50 · 2,407 ms TTFT p50 · 6,655 ms p90
| Model | Share | Tokens 7d | Requests | Cache | TTFT p50 | tok/s | Monthly limit |
|---|---|---|---|---|---|---|---|
| DeepSeek V4.1 Flash | 85.1% | 1.53B | 7,879 | 92.4% | 2,414 ms | 165 | $80.00/mo |
| MiMo v2.6 Flash | 14.9% | 267M | 2,753 | 66.7% | 2,357 ms | 77 | $60.00/mo |
Z.ai Coding Plan - Max (Legacy v2)
$56.00/mo5h 12.0% · weekly 67.0% · resets 2026-11-02 04:43Z
- Capacity
- 921M per 7d · weekly window
- $/B
- $14.20
- Traffic, last 7 days
- 664M tokens · 4,801 requests · cache 94.1% · 7 days
- Speed
- 76 tok/s p50 · 5,064 ms TTFT p50 · 9,972 ms p90
| Model | Share | Tokens 7d | Requests | Cache | TTFT p50 | tok/s | Monthly limit |
|---|---|---|---|---|---|---|---|
| GLM-5.3 (no list price) | 81.8% | 543M | 3,073 | 96.7% | 4,809 ms | 80 | N/Ano per-model limit |
| GLM-5.3-Flash | 18.2% | 121M | 1,728 | 80.8% | 6,098 ms | 53 | N/Ano per-model limit |
QwenCloud - Token Plan Individual
$10.00/momonthly 57.9% · resets 2026-11-24 16:00Z
- Capacity
- 452M per 30d · monthly window
- $/B
- $22.12
- Traffic, last 7 days
- 475M tokens · 2,654 requests · cache 93.4% · 7 days
- Speed
- 148 tok/s p50 · 3,867 ms TTFT p50 · 9,349 ms p90
| Model | Share | Tokens 7d | Requests | Cache | TTFT p50 | tok/s | Monthly limit |
|---|---|---|---|---|---|---|---|
| DeepSeek V4.1 Flash (1017) | 59.4% | 282M | 1,755 | 96.1% | 3,218 ms | 155 | N/Ano per-model limit |
| DeepSeek V4.1 Flash | 27.9% | 133M | 622 | 87.8% | 5,835 ms | 155 | N/Ano per-model limit |
| Qwen3.8 Flash | 7.2% | 34.0M | 142 | 93.6% | 5,819 ms | 72 | N/Ano per-model limit |
| Qwen3.8 Max | 5.5% | 26.3M | 135 | 93.3% | 7,236 ms | 40 | N/Ano per-model limit |
Formulas, sources and limits: /methodology