Follow the price.
Understand the timing.
Explore the dated price-history entries recorded with the model catalogue. A line connects observations; it does not imply daily measurements.
56
Models Tracked
$0.10
Cheapest
$30
Most Expensive
17
Providers
Click to add/remove models from the chart:
Recorded price changes
| Recorded date | Input / 1M tokens | Output / 1M tokens |
|---|---|---|
| 2024-05-13 | $5.00 | $15.00 |
| 2024-10-01 | $2.50 | $10.00 |
These are the price changes recorded against the model in the catalogue, with the date each one was recorded. Sub-cent rates keep four decimal places, which is the difference between $0.0028 and $0.00.
Average Price by Provider
Optimize Your AI Costs
Swfte Connect routes each request to the optimal model based on complexity: saving 30-60% on API costs without sacrificing quality.
AI Model Pricing: A Race to the Bottom
AI API pricing has undergone a dramatic transformation. When GPT-4 launched in March 2023, it cost $30 per million input tokens. Today, comparable-quality models are available for under $1/M: a 97% reduction in just three years.
The price war is driven by competition on three fronts: open-source models (Meta Llama, DeepSeek, Qwen), aggressive pricing from Google (Gemini Flash at $0.10/M), and continuous optimization by all providers. This page tracks 56 models across 17 providers, updated every 6 hours.
How to Use Pricing Trends
- Track price drops: When providers cut prices (often triggered by competitor launches), it shows up here within hours
- Compare providers: Use the provider filter to see all models from one provider and compare their pricing tiers
- Evaluate value: Check the leaderboard alongside trends to find the best quality-per-dollar
- Plan migrations: Use the cost calculator to estimate savings from switching models