I was checking the latest competitive data on AI models. ChatGPT is still number one. Most widely used. Highest monthly active users.
But something shifted in 2025.
Google’s Gemini, which started the year outside the top 10, climbed steadily and is now holding the second spot. Claude has been battling between second and third position. Perplexity is in third.
What’s interesting isn’t who’s first. It’s that the gap is closing fast.
In late 2024, ChatGPT had such dominance that competitors seemed irrelevant. OpenAI had the ecosystem. The integrations. The brand. The first-mover advantage.
But by December 2025, something changed. Three things, actually.
First, Claude got really good at coding. Like, absurdly good. Enterprises started using Claude specifically for software development because it outperformed ChatGPT for that use case. One company told me they tried GPT-5 for coding but switched back to Claude for a specific project because Claude just worked better.
Second, Gemini became cheap. Not just cheaper. Actually competitive on price while maintaining quality. Google integrated it deeply into their ecosystem i.e. Workspace, Gmail, Search. If you’re already in the Google world, it makes sense to use Gemini.
Third, OpenAI became focused on ChatGPT the product. ChatGPT the interface. ChatGPT the consumer brand. But enterprises want APIs. They want flexibility. They want options.
So now OpenAI is fighting two enemies. The first enemy is competitors who are better specialized (Claude for code, Gemini for integration). The second enemy is its own brand. ChatGPT is so consumer-focused that enterprises sometimes choose competitors just for the image of “enterprise-grade AI.”
OpenAI keeps releasing new models - GPT-5, then GPT-5.1, then GPT-5.2. Each one claims to be better. But does market share actually move? Not much.
The war is moving from “who has the best model” to “who can embed themselves deepest in your workflow.”
And that’s a different game than the one OpenAI won at launch.