Hey Vapi Team, we ran into a possible bug with LLM cost calculation on a short test call. Call Details: Call ID: 7f30a6e8-26b3-457f-ad04-80915c0d9fd5 Assistant ID: f7cbe813-eda4-4b24-9e96-8ad2051fcad4 (Strom & Gas) Timestamp: 2025-08-28 23:52 UTC Expected outcome: GPT-4o Cluster cost ~$0.05/min → <$0.05 for 54s Actual outcome: LLM cost was $5.99 for 54s (total $6.07) Token usage: 1,832,937 prompt tokens / 82 completion tokens This number of prompt tokens seems highly implausible for such a short call and likely indicates a calculation or logging error. I’ve attached screenshots for: – Cost breakdown – LLM breakdown – Call log Can someone look into this? Our account is now out of credit because of this single test. Thanks!

Hey, thanks for sharing your details. Our team will look into it and get back to you soon with an update.

any update?

\[ { "type": "transcriber", "cost": 0.01, "minutes": 0.9887166667, "transcriber": { "model": "nova-2", "provider": "deepgram" } }, { "type": "model", "cost": 5.9, "promptTokens": 1832937, "completionTokens": 82, "model": { "model": "gpt-4o", "provider": "openai" } }, { "type": "voice", "cost": 0.01, "characters": 370, "voice": { "model": "eleven_flash_v2_5", "voiceId": "BIvP0GN1cAtSRTxNHnWS", "provider": "11labs" } }, { "type": "vapi", "cost": 0.05, "minutes": 0.9037, "subType": "normal" }, { "type": "analysis", "cost": 0.000132, "analysisType": "summary", "promptTokens": 340, "completionTokens": 12, "model": { "model": "gemini-2.5-flash", "provider": "google" } }, { "type": "analysis", "cost": 0.0006256, "analysisType": "successEvaluation", "promptTokens": 2002, "completionTokens": 10, "model": { "model": "gemini-2.5-flash", "provider": "google" } }, { "type": "knowledge-base", "cost": 0, "promptTokens": 0, "completionTokens": 0, "model": { "model": "gemini-1.5-flash", "provider": "google" } } \]