I have an agent that uses two AI models: Gemini 2.0 Flash (default fast LLM, fallback LLM) and GPT-4o mini (default best LLM, autonomous language model, RAG language model). ## 1. Reference API Pricing According to the official API documentation, the costs are: * Gemini 2.0 Flash: $0.10 / M input tokens | $0.40 / M output tokens. * GPT-4o mini: Starting at $0.15 / M input tokens | $0.60 / M output tokens. ## 2. Activity Summary (January 19) On January 19, the recorded LLM activity was: * Input tokens: 737,403 ($0.737403$ M) * Output tokens: 16,738 ($0.016738$ M) Reported Daily Cost: $3,389.00 USD ## 3. Theoretical Cost Calculation When calculating the cost based on the official API rates, the results are significantly lower than the actual charge: ### Gemini 2.0 Flash Calculation: * Input cost: $0.737403 \times 0.10 = \$0.0737$ * Output cost: $0.016738 \times 0.40 = \$0.0067$ * Total: $0.0804 USD ### GPT-4o mini Calculation: * Input cost: $0.737403 \times 0.15 = \$0.1106$ * Output cost: $0.016738 \times 0.60 = \$0.0100$ * Total: $0.1206 USD ## 4. Contextual Notes * I am not using any personality, knowledge, translation, vision, policy, or analytics agents. * The difference between the calculated price (~$0.12) and the charged price ($3,389.00) is approximately 28,000x higher than expected. * Screenshots of the dashboard are attached for reference. https://cdn.discordapp.com/attachments/1463252671640240139/1463252673003520183/Screenshot_2026-01-20_at_14.16.41.png?ex=697b0aff&is=6979b97f&hm=aed5dc3f0cebc0ef5afb37f78373da835b29dd6c8d7cb8421a4b731b42ba05a5& https://cdn.discordapp.com/attachments/1463252671640240139/1463252673515229471/Screenshot_2026-01-20_at_14.17.59.png?ex=697b0aff&is=6979b97f&hm=5f4d08c7d197394592e083ea4c1735ae62bf9e583919fd5c960c46756403c8ed& https://cdn.discordapp.com/attachments/1463252671640240139/1463252674223931442/Screenshot_2026-01-20_at_14-08-54_AMA_-_Produccion_-_Botpress_Studio.png?ex=697b0aff&is=6979b97f&hm=647f166c03ad3cddaa6c19b7c386f72518a58cf2ab43f898509ca92bb78381d2&

Hey @wonderful-monitor-65437 as a plus plan member you have direct access to our solution engineers! I urge you to reach out and ask there :)O