forked from kaiban-ai/KaibanJS
-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Adds cost calculations per task and workflow.
- Loading branch information
1 parent
d7f53ef
commit 3026d74
Showing
6 changed files
with
272 additions
and
11 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,140 @@ | ||
| import {logger} from "./logger"; | ||
|
|
||
| const modelsPricing = [ | ||
| { | ||
| "modelCode": "gpt-4o-mini", | ||
| "provider": "openai", | ||
| "inputPricePerMillionTokens": 0.150, | ||
| "outputPricePerMillionTokens": 0.600, | ||
| "features": "Most cost-efficient with vision capabilities" | ||
| }, | ||
| { | ||
| "modelCode": "gpt-3.5-turbo", | ||
| "provider": "openai", | ||
| "inputPricePerMillionTokens": 0.50, | ||
| "outputPricePerMillionTokens": 1.50, | ||
| "features": "Cost-effective option" | ||
| }, | ||
| { | ||
| "modelCode": "gpt-3.5-turbo-0125", | ||
| "provider": "openai", | ||
| "inputPricePerMillionTokens": 0.50, | ||
| "outputPricePerMillionTokens": 1.50, | ||
| "features": "Cost-effective option" | ||
| }, | ||
| { | ||
| "modelCode": "gpt-4o", | ||
| "provider": "openai", | ||
| "inputPricePerMillionTokens": 5.00, | ||
| "outputPricePerMillionTokens": 15.00, | ||
| "features": "Enhanced multimodal capabilities" | ||
| }, | ||
| { | ||
| "modelCode": "gpt-4-turbo", | ||
| "provider": "openai", | ||
| "inputPricePerMillionTokens": 10.00, | ||
| "outputPricePerMillionTokens": 30.00, | ||
| "features": "High speed and power" | ||
| }, | ||
| { | ||
| "modelCode": "gpt-4", | ||
| "provider": "openai", | ||
| "inputPricePerMillionTokens": 30.00, | ||
| "outputPricePerMillionTokens": 60.00, | ||
| "features": "Latest generation AI" | ||
| } | ||
| ]; | ||
|
|
||
| /** | ||
| * Calculates the approximate cost of using a specified AI model based on the number of input and output tokens. | ||
| * If the model code does not match any in the predefined list, it returns a standardized error response with costs set to -1. | ||
| * | ||
| * @param {string} modelCode - The unique code identifier for the AI model. | ||
| * @param {object} llmUsageStats - An object containing usage statistics including: | ||
| * - inputTokens: The number of input tokens processed. | ||
| * - outputTokens: The number of output tokens generated. | ||
| * - callsCount: The total number of calls made. | ||
| * - callsErrorCount: The number of failed calls. | ||
| * - parsingErrors: The number of errors encountered during parsing. | ||
| * @returns {object} An object containing: | ||
| * - costInputTokens: The calculated cost for the input tokens or -1 if the model is not found. | ||
| * - costOutputTokens: The calculated cost for the output tokens or -1 if the model is not found. | ||
| * - totalCost: The total cost combining input and output tokens or -1 if the model is not found. | ||
| * | ||
| * @example | ||
| * // Calculate costs for a known model with usage stats | ||
| * const llmUsageStats = { | ||
| * inputTokens: 500000, | ||
| * outputTokens: 200000, | ||
| * callsCount: 10, | ||
| * callsErrorCount: 1, | ||
| * parsingErrors: 0 | ||
| * }; | ||
| * const costDetails = calculateCost('gpt-4o-mini', llmUsageStats); | ||
| * console.log(costDetails); | ||
| */ | ||
| function calculateTaskCost(modelCode, llmUsageStats) { | ||
| const model = modelsPricing.find(m => m.modelCode === modelCode); | ||
| if (!model) { | ||
| return { | ||
| costInputTokens: -1, | ||
| costOutputTokens: -1, | ||
| totalCost: -1 | ||
| }; | ||
| } | ||
|
|
||
| const inputCost = (llmUsageStats.inputTokens / 1000000) * model.inputPricePerMillionTokens; | ||
| const outputCost = (llmUsageStats.outputTokens / 1000000) * model.outputPricePerMillionTokens; | ||
| const totalCost = inputCost + outputCost; | ||
|
|
||
| return { | ||
| costInputTokens: parseFloat(inputCost.toFixed(4)), | ||
| costOutputTokens: parseFloat(outputCost.toFixed(4)), | ||
| totalCost: parseFloat(totalCost.toFixed(4)) | ||
| }; | ||
| } | ||
|
|
||
|
|
||
| function calculateTotalWorkflowCost(modelUsageStats) { | ||
| let totalInputCost = 0; | ||
| let totalOutputCost = 0; | ||
| let totalCost = 0; | ||
| let totalInputTokens = 0; | ||
| let totalOutputTokens = 0; | ||
| let allPricesAvailable = true; | ||
|
|
||
| Object.keys(modelUsageStats).forEach(modelCode => { | ||
| const stats = modelUsageStats[modelCode]; | ||
| const model = modelsPricing.find(m => m.modelCode === modelCode); | ||
| if (!model) { | ||
| logger.warn(`No pricing information found for model ${modelCode}`); | ||
| allPricesAvailable = false; | ||
| return; | ||
| } | ||
| const inputCost = (stats.inputTokens / 1000000) * model.inputPricePerMillionTokens; | ||
| const outputCost = (stats.outputTokens / 1000000) * model.outputPricePerMillionTokens; | ||
| totalInputCost += inputCost; | ||
| totalOutputCost += outputCost; | ||
| totalCost += inputCost + outputCost; | ||
| totalInputTokens += stats.inputTokens; | ||
| totalOutputTokens += stats.outputTokens; | ||
| }); | ||
|
|
||
| if (!allPricesAvailable) { | ||
| return { | ||
| costInputTokens: -1, | ||
| costOutputTokens: -1, | ||
| totalCost: -1 | ||
| }; | ||
| } | ||
|
|
||
| return { | ||
| costInputTokens: parseFloat(totalInputCost.toFixed(4)), | ||
| costOutputTokens: parseFloat(totalOutputCost.toFixed(4)), | ||
| totalCost: parseFloat(totalCost.toFixed(4)) | ||
| }; | ||
| } | ||
|
|
||
| export { calculateTaskCost, calculateTotalWorkflowCost }; | ||
|
|
||
|
|
Oops, something went wrong.