The o3 model is now 80% cheaper same capabilities, much lower cost. Perfect for coding, tool use, and complex agent workflows.
New Pricing:
- Input: $2.00 / 1M tokens
- Cached input: $0.50 / 1M tokens
- Output: $8.00 / 1M tokens
This now matches GPT-4.1 pricing.
I’m planning to use it for some of my heavier AI automations.
Anyone here tested o3 lately? How does it perform for you?