Unverified50% confidenceFactExact time
Token consumption is paused while the watchdog waits—it is only polling API status and incurs no additional cost; failed retries, by contrast, may consume more tokens because the context is sent repeatedly.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/12/2026
First Seen
Valid until: 1/10/2027
Sources
Related Entities
Related Claims
UnverifiedWhen a streaming response is interrupted midway, the portion of output tokens already generated is still billed, even though the client may not have received the complete result.73% similarUnverifiedReasoning content is often billed the same as regular text in terms of token counting, so retaining reasoning increases the number of tokens carried in requests and affects call costs.66% similarUnverifiedPaid AI workspaces typically restrict write scope via OAuth Apps or Fine-grained Tokens, and these tokens become invalid after the subscription is cancelled.65% similarUnverifiedEach retry resends the complete prompt input tokens and is billed in full; a request with 4,000 input tokens retried 3 times will consume 16,000 tokens.65% similarUnverifiedIn frameworks where prompt caching is already enabled, the unit price for cache hits is only 1/50 that of new input, so cost savings are limited when the volume of judgments is not large.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1009979API
curl https://kongchang.com/api/v1/knowledge/claims/1009979MCP
get_claim(id=1009979)