OpenRouter Data Shows Claude Opus 4.8 Dominating AI Model Spending

Curious which AI models are actually pulling in the money? Fresh OpenRouter data from the last 30 days breaks down spending by task — and the leaderboard is not what you'd expect.

What Happened

New OpenRouter spend data for the last ~30 days shows Claude Opus 4.8 dominating 17 out of 29 categories by dollars spent. In Research & Reports alone, Opus 4.8 takes 45% of the spend.

Customer Support is far more competitive. Gemini 3 Flash leads that category with only 7.4%, meaning a lot of models are splitting the rest.

There is also a key gap: spend does not equal tokens. Ranking models by token usage instead of dollars gives a very different leaderboard.

Why It Matters

Anthropic is ahead overall, but the gap is closing in some of the biggest categories. If you are choosing a model for cost or performance, dollars spent and tokens used can tell different stories — worth checking OpenRouter's data before you commit.