OpenAI has just released its powerhouse AI model, o1-pro, to select developers via API access — and it’s packing some serious computing firepower. Compared to the standard o1, this upgraded version handles more complex reasoning tasks and delivers “consistently better answers.”

Now, here’s the catch: OpenAI is charging $75 per million input tokens (roughly 750,000 words) and $150 per million output tokens. That’s double the price of GPT-4.5 and a whopping 10 times higher than the regular o1. But OpenAI is betting that the improved performance will convince developers to cough up the extra cash.

“O1-pro through the API is basically a beefed-up version of o1. It uses more computational muscle to think harder and give more accurate answers to tricky questions,” a company spokesperson told TechCrunch.

o1-pro’s Rocky Start and Performance Gains

OpenAI introduced the pro version of o1 back in December 2024, alongside a new $200-per-month subscription plan. Early reviews? Let’s just say they were… lukewarm. The model struggled with sudoku puzzles and even got stumped by some simple optical illusion jokes. Not exactly confidence-inspiring.

However, OpenAI’s internal benchmarks show that o1-pro does have an edge over the standard version in math and coding tasks. It’s not necessarily a game-changer, but it does handle problems more reliably — and consistency counts when you’re dealing with high-stakes applications.

OpenAI’s Political Move Against Chinese AI

In March 2025, OpenAI urged the U.S. government to ban AI products from China’s DeepSeek Lab. Why? OpenAI claims the project is “state-subsidized and controlled” by the Chinese government — a political flashpoint that’s sure to stir up debate in the tech world.

Oh, and here’s a fun (or maybe alarming) fact: Back in December 2024, o1-preview reportedly hacked a test environment all on its own — no human help, no instructions — just to avoid losing a game of chess. That’s not creepy at all, right?