Weekly Usage Report — Week 1 (Apr 6–12): 97 Million Accounted Tokens for £9.24
Week 1: 51.8M visible tokens plus 45.6M cached tokens, for 97.4M total accounted Hermes tokens across 88 sessions. Opus-equivalent API cost: about £1,188.
This is Week 1 of an ongoing series. Every Monday, I pull back the curtain on what our AI agent actually does — and what it actually costs. No marketing fluff. Just honest numbers from our Mission Control dashboard.
Token accounting
This report separates visible prompt/completion tokens from cached context. Visible tokens show fresh input/output work; cached tokens show repeated context reused during long agent sessions. Together, they show the full model-traffic footprint for the week.
- Visible tokens (input + output): 51,849,431 (51.8M)
- Cached tokens (cache-read/write): 45,585,280 (45.6M)
- Total accounted tokens: 97,434,711 (97.4M)
- Sessions: 88
- Input tokens: 51,458,049
- Output tokens: 391,382
- Total cost: £9.24/week
- Opus-equivalent API cost: approximately £1,188
The week in one picture
This is the headline version of Week 1: 51.8M tokens, 88 sessions, £9.24 in subscription route cost — and a brutal comparison with what the same partial-week workload would cost on per-token pricing.
Top visible model routes
| Model | Type | Share of visible route tokens | Cost |
|---|---|---|---|
| GLM-5.1 | Cloud (OAuth) | ~49% | £4.62/wk |
| Qwen 3.5 9B | Local (Ollama) | ~25% | Free |
| GPT-5.3 Codex | Cloud (OAuth) | ~25% | £4.62/wk |
These shares describe the visible input/output route mix only. Cached context is included in the week’s total accounted-token figure above, but it is not allocated cleanly by model route here. The local model still carried a meaningful slice of the visible workload at zero marginal cost.
The Price Comparison That Should Make You Angry
What would 52M tokens cost on per-token pricing?
- Claude Opus 4.6: £641 → 69x my cost
- Claude Sonnet 4: £103 → 11x
- GPT-5.3 Codex (per-token): £172 → 19x
- DeepSeek Chat: £11 → 1x
On Opus per-token pricing, this partial week would cost £641. For two days of AI usage by one person.
I paid £9.24.
Daily Breakdown
- Mon Apr 6: 0 sessions, 0 visible + 0 cached = 0 total accounted tokens, 0.0% of the week. Work note: No tracking yet.
- Tue Apr 7: 0 sessions, 0 visible + 0 cached = 0 total accounted tokens, 0.0% of the week. Work note: No tracking yet.
- Wed Apr 8: 0 sessions, 0 visible + 0 cached = 0 total accounted tokens, 0.0% of the week. Work note: No tracking yet.
- Thu Apr 9: 0 sessions, 0 visible + 0 cached = 0 total accounted tokens, 0.0% of the week. Work note: No tracking yet.
- Fri Apr 10: 0 sessions, 0 visible + 0 cached = 0 total accounted tokens, 0.0% of the week. Work note: No tracking yet.
- Sat Apr 11: 20 sessions, 22,130,569 visible (22.1M) + 21,345,408 cached (21.3M) = 43,475,977 total accounted tokens (43.5M), 44.6% of the week; cache share 49.1%, visible share 50.9%. Work note: Early setup — model evaluation, local testing.
- Sun Apr 12: 68 sessions, 29,718,862 visible (29.7M) + 24,239,872 cached (24.2M) = 53,958,734 total accounted tokens (54.0M), 55.4% of the week; cache share 44.9%, visible share 55.1%. Work note: School data research, Mission Control styling, web crawling.
The Stack
| Component | Cost | Type |
|---|---|---|
| GLM-5.1 (cloud) | £4.62/wk | OAuth subscription |
| GPT-5.3 Codex (cloud) | £4.62/wk | OAuth subscription |
| Qwen 3.5 9B (local) | £0 | Local Ollama |
| Gemma 4 31B (cloud) | £0 | Free tier |
| MiniMax M2.7 (cloud) | £0 | Free tier |
| Total | £9.24/wk | £480/year |
No API keys. No per-token billing. No surprise invoices.
💡 Weekly Usage Report #1. Published every Monday with the previous week's (Mon–Sun) data. Numbers from the Mission Control token tracking dashboard.
Found this useful? 👉 Follow @Raf_VRS for more transparent AI insights 👉 Support the work: ko-fi.com/rafvrs #ModelBenchmarking #TokenUsage #AIAgents #HardInterference