02
Cost by model
Contribution to the current estimate
Estimated monthly cost
$256.96/ month
Actual Hermes usage · last 30 days
Total tokens455.02M31.66M input · 2.19M output
Daily estimate$8.57monthly ÷ 30
Annual estimate$3,083.54monthly × 12
Cached / other92.56%421.16M tokens
Edits save automatically in this browser.
01
Model calculation
Token quantities and per-million USD prices are editable. Calculations update immediately.
| Model | Input tokens uncached | Input $/M | Output tokens | Output $/M | Cached tokens assumed billable reads | Cache $/M assumption | Subtotal | Remove |
|---|---|---|---|---|---|---|---|---|
| Combined estimate USD base | $256.96 | |||||||
Accepted token formats: 455,019,030, 455m, 500k, 1.5b. Cache rates are editable assumptions unless Nous Portal explicitly lists them.
01
Simple blended estimate
Split one monthly token total across input, output, and cached usage.
03
Usage by profile
Fixed snapshot from Hermes insights · last 30 days
fusion GPT-5.6 Sol257.66M56.63%
full GPT-5.5193.20M42.46%
telegram-lean GPT-5.53.92M0.86%
default GPT-5.6 Sol229.47K0.05%
fable No usage00%
04
Assumptions & methodology
Formula
Σ(tokens ÷ 1,000,000 × price per 1M)Input, output, and cached tokens are priced separately for every model row. Simple mode uses a user-defined percentage split.
Pricing provenance
Nous Portal catalog: GPT-5.6 Sol (batch) $2 input / $12 output; GPT-5.5 $4 input / $24 output per 1M. Checked Aug 9, 2026.
Assumption: cache-read defaults are 10% of input ($0.20 and $0.40 per 1M). They are not represented as published Nous Portal rates.
Estimate caveat
This client-side estimate is not an invoice or guaranteed Portal charge. “Cached/other” is the residual between Hermes total tokens and uncached input plus output, treated here as cache reads for estimation.
Currency conversions are illustrative fixed display rates; USD remains authoritative.