Document ProcessingGPT-3.5 Turbo

Invoice Data Extractor Using GPT-3.5 Turbo

Optimized prompts and cost breakdowns for running invoice data extractor with GPT-3.5 Turbo.

About GPT-3.5 Turbo

Provider: openai
Context Window: 16K tokens
Input Cost: $0.5 / 1M tokens
Output Cost: $1.5 / 1M tokens

Best For

Legacy supportSimple classificationQuick responses

Optimized Prompts for GPT-3.5 Turbo

These prompts are tailored to work best with GPT-3.5 Turbo's capabilities

extraction

Extract the following from this invoice:
- Vendor name and address
- Invoice number and date
- Due date
- Line items (description, quantity, unit price, total)
- Subtotal, tax, and total

Return as JSON.

{{invoice_text}}

Cost Breakdown

Estimated costs for running this workflow with GPT-3.5 Turbo

input tokens: ~1278 tokens
output tokens: ~209 tokens
cost per request: $0.0467
monthly 1000 requests: $50.00

Rate Limiting Considerations

Openai has rate limits of approximately 9314 RPM. Consider implementing request queuing for high-volume usage.

Monitoring with Administrate

Track GPT-3.5 Turbo API latency and error rates. Set up alerts for rate limit warnings and cost thresholds.