DSH Cost Estimate

Key features

• Pre-estimationShow prestimated input token, output token and cost ranges before the model is generated
• Prefer using username (inputTokens / cacheReadTokens / carcheWriteTokens) for precise anchoring of each request
• Pure client realizationEstimates and renderings are done entirely on the Web Client (Conversion NodeDefinion + transfer.chat.node slot).
• The estimate is for information purposes only and is based on the actual DeepSeek deductions.
• After reconciliationafter the answer (takes the provider real usage), the same line flips to actual token and cost, and displays a cache hit rate