command: /tokens
is_error: false
message:
Token Usage:
 ─────────────────────────────
 Active context:        ~48 / 1000000 (0.0%)
 Last API input:        100 (turn telemetry; may count repeated prefix across tool rounds)
 Last API output:       25
 Cache hit/miss:        70 hit / 30 miss (telemetry/cost only)
 Cumulative tokens:     1234 (session usage telemetry)
 Priced amount:         $0.0500
 API messages:          0
 Chat messages:         0
 Model:                 deepseek-v4-pro
Cache-write (cache-creation) tokens this session: not reported. Billed above the input rate where the provider publishes a write premium; see /cache for the per-turn breakdown.

This is an estimate, not a bill: it is computed locally from provider-reported token usage and published rates, and can differ from your invoice.
Covered: 1 of 1 money-metered turn(s) priced.
action: <none>
