| Usage Level | GPT-3.5 Turbo | GPT-4o Mini | Savings |
|---|---|---|---|
| Light (100 calls/day) | $1.65 | $0.59 | Save $1.07 with GPT-4o Mini |
| Medium (1K calls/day) | $30.00 | $10.80 | Save $19.20 with GPT-4o Mini |
| Heavy (10K calls/day) | $375.00 | $135.00 | Save $240.00 with GPT-4o Mini |
| Enterprise (100K/day) | $4,500.00 | $1,620.00 | Save $2,880.00 with GPT-4o Mini |
GPT-4o Mini is 63% cheaper overall. GPT-3.5 Turbo costs $0.5/M input vs GPT-4o Mini at $0.15/M input.
At 800 input + 400 output tokens per call: GPT-3.5 Turbo costs $10.00 and GPT-4o Mini costs $3.60.
With a proxy like Tokonomics, switching is a one-line config change. Route different features to different models and track costs per model automatically.