GPT-4o API Inference Cost Measure

From GM-RKB
Jump to navigation Jump to search

A GPT-4o API Inference Cost Measure is a 3rd-Party LLM inference cost measure for the cost associated with generating each output token during an inference API call of GPT-4o.



References

2024

  • https://artificialanalysis.ai/models/gpt-4o#pricing
    • NOTES:
      1. Blended Price**: GPT-4o has a blended price of $7.50 per 1M tokens, combining both input and output token costs.
      2. Input Token Price**: The input token price for GPT-4o is $5.00 per 1M tokens.
      3. Output Token Price**: The output token price for GPT-4o is $15.00 per 1M tokens.
      4. Cost-Effective Quadrant**: In the quality vs. price analysis, GPT-4o is in the top-left quadrant, indicating high quality at a relatively lower price.
      5. Balanced Pricing Structure**: GPT-4o shows a balanced structure in input and output prices, making it cost-efficient.
      6. Competitive Performance**: GPT-4o offers a high-quality index and efficient performance metrics at a moderate price point.
      7. High Output Speed**: With an output speed of 83 tokens per second, GPT-4o is efficient in generating tokens quickly.
      8. Low Latency**: GPT-4o has a low latency of 0.44 seconds, beneficial for real-time applications.
      9. Comparison Advantage**: GPT-4o is more affordable than high-cost models like Claude 3 Opus while maintaining competitive performance.
      10. Attractive Option**: The combination of reasonable pricing, high output speed, and low latency makes GPT-4o an appealing choice for various AI applications.