Skip to main content
SCAIRE

Token Cost

Cost and currency controls for token tracking, plus the per-model rate table and calibration below.

Token Pricing Rates

Per-model pricing per 1M tokens (USD). These rates power cost tracking on the Statistics page and chat badges. Rates are seeded from public provider pricing and can be edited freely.

Model Input / 1M Output / 1M Cached / 1M Factor Status Actions
glm-5.2 $1.40 $4.40 $0.26 1.00 Active
glm-5.1 $1.40 $4.40 $0.26 1.00 Active
glm-5 $1.00 $3.20 $0.20 1.00 Active
glm-5-turbo $1.20 $4.00 $0.24 4.99 Active
glm-4.7 $0.60 $2.20 $0.11 1.00 Active
glm-4.7-flashx $0.07 $0.40 $0.01 1.00 Active
glm-4.7-flash $0.00 $0.00 $0.00 1.00 Active
glm-4.6 $0.60 $2.20 $0.11 1.00 Active
glm-4.5 $0.60 $2.20 $0.11 1.00 Active
glm-4.5-x $2.20 $8.90 $0.45 1.00 Active
glm-4.5-airx $1.10 $4.50 $0.22 1.00 Active
glm-4.5-air $0.20 $1.10 $0.03 1.00 Active
glm-4.5-flash $0.00 $0.00 $0.00 1.00 Active
glm-4-32b-0414-128k $0.10 $0.10 $0.00 1.00 Active

Token Cost

Cost & Currency

Calibrate first while ON, then turn OFF

ISO 4217 code (usd, cad, eur, etc.)

Cost Coverage

Whether your real usage is priced, and what it costs.

Rate coverage
33%
1/3 used models priced
Spend (30 days)
$0.00
$0.00 in the last 7 days
Rates on file
14
14 active across 1 provider
Coverage
2 model(s) used in the last 30 days have no rate. Their cost is not tracked.
z-ai/glm-5.2 deepseek-ai/deepseek-v4-pro

Add a rate for each unpriced model in the table to make cost tracking complete.

Calibration
Correction factors set 1

Calibration refines the local token estimate against the provider's real counts. A factor of 1.0 means uncalibrated.

Why Add Rates
  • Track what each diagnostic run actually costs in tokens and money.
  • Per-model rates give accurate cost attribution, not a global average.
  • The Statistics page can only show cost when rates are set.
No rates means token usage is tracked but cost is silent. Seed curated rates or add your own.
Quick Links
Calibrate

Sends short probe messages to the default provider and measures the actual vs estimated token ratio. Uses real tokens.