Skip to main content

OpenAI Pricing

GPT-4, GPT-4o, and GPT-3.5 Turbo pricing for CloudCompare integration.

Model Pricing

GPT-4o

Flagship multimodal model with vision and audio capabilities.

TierInput (per 1M tokens)Output (per 1M tokens)
Standard$2.50$10.00
Batch API (async)$1.25$5.00

GPT-4o-mini

Cost-efficient small model with strong performance.

TierInput (per 1M tokens)Output (per 1M tokens)
Standard$0.15$0.60
Batch API$0.075$0.30

GPT-4 Turbo

Previous generation flagship.

TierInput (per 1M tokens)Output (per 1M tokens)
Standard$10.00$30.00
Batch API$5.00$15.00

GPT-3.5 Turbo

Budget option for simple tasks.

TierInput (per 1M tokens)Output (per 1M tokens)
0125 version$0.50$1.50

Vision Pricing

GPT-4o vision analysis included in token count.

Image SizeCost Equivalent
< 512x512~85 tokens
512x512 to 1024x1024~170 tokens
> 1024x1024~510 tokens

Free Tier

BenefitAmount
Monthly tokens100k (GPT-4o-mini only)
Rate limit3 RPM
ValidityForever

Cost Examples for Point Cloud Processing

Point Cloud Analysis (1MB scan)

OperationTokens UsedCost
Scene description~2,000 input + 500 output~$0.008
Classification rules~1,000 input + 800 output~$0.005
Full report generation~3,000 input + 2,000 output~$0.026

Monthly Cost Scenarios

Use CaseRequests/monthAvg tokensMonthly Cost
Light analysis (10 scans)303,000~$0.90
Medium (50 scans)1505,000~$4.50
Heavy (200 scans)6008,000~$18.00

API Rate Limits

PlanRPMRPDTPM
Free3-100k
Pay-as-you-go500-5M
Enterprise10,000+UnlimitedUnlimited

CloudCompare Configuration

{
"openai": {
"api_key": "sk-...",
"organization": "org-...",
"default_model": "gpt-4o",
"max_tokens": 4096,
"temperature": 0.3
}
}

Comparison with Other Vendors

MetricOpenAIAnthropicGoogle
Best forVersatilityReasoningLong context
Cheapest model$0.15/1M in$0.25/1M in$0.075/1M in
Context window128k200k2M
VisionYesYesYes

Tips to Reduce Costs

  1. Use GPT-4o-mini for simple classification tasks
  2. Batch requests for non-real-time processing (50% discount)
  3. Cache responses for repeated queries
  4. Reduce image resolution for point cloud renders
  5. Use prompts efficiently - include only necessary context