Claude AI's New Opus 4.7 Uses 35% More Tokens Than Previous Version
Anthropic's new Claude Opus 4.7 AI model uses up to 35% more tokens than its predecessor for the same input text. The increase comes from a new tokenizer system that breaks down text differently.
Anthropic released Claude Opus 4.7 with what the company called unchanged pricing, but users are discovering their actual costs are rising sharply. The new version uses a different tokenizer that can create up to 35% more tokens from identical input text.
Tokens are how AI companies measure and charge for text processing. Think of them like units of text - more tokens mean higher bills. While Anthropic kept the price per token the same, users now need more tokens for the same work.
The new model includes other changes like better vision capabilities, improved instruction following, and new self-verification features. But some users on Reddit are complaining about bugs and extended processing times, with one reporting their task running for over an hour.
The pricing structure means customers face a hidden cost increase despite the company's claims of stable pricing. Real-world usage costs could rise substantially depending on how much text users process.
If you use Claude AI for work or projects, your bills could jump significantly even though the company kept the same per-token price. More tokens mean higher costs for the same amount of text you process.
Users will likely pressure Anthropic to address the token increase or adjust pricing to reflect actual costs.
Was this article helpful?
0 people found this helpful