Zhipu GLM-5 Raises Prices by 83%: Chinese LLMs Finally Dare to Talk Money

Honestly, when I saw the news that Zhipu GLM-5-Turbo raised prices by 83%, my first reaction was: Finally, someone dared to do this.

You know what the Chinese LLM market has been like for the past two years?

“We’re 90% cheaper than OpenAI!”

“New users get 10 million free tokens!”

“API prices cut another 50%!”

The entire industry has been competing like a vegetable market, as if higher prices mean you’re the fool. But the problem is — even if your price is low, if the model is bad, developers will still vote with their feet.

Where does Zhipu’s confidence in this price increase come from?

I checked the technical report. GLM-5’s improvement in programming capability is indeed impressive. Some programmers tested it and said it’s close to Claude Opus 4.5’s level — if this is true, then this price increase is completely reasonable. After all, everyone knows what level Claude Opus’s API pricing is at.

But there’s a more interesting phenomenon here: Chinese LLMs seem to be experiencing a turning point from “competing on price” to “competing on value.”

Before, everyone competed on who was cheaper. Now they’re starting to compete on who can solve problems better. This shift itself is more worth discussing than the price increase.

My personal experience with domestic models is — daily conversations and simple tasks are indeed sufficient, but when it comes to complex programming, long-text reasoning and other hardcore scenarios, I still have to switch back to Claude or GPT. If GLM-5 can truly catch up at this level, then the price increase is market behavior, not harvesting韭菜.

Of course, there are risks too.

The price increase means user expectations are raised. Before, “cheap but mediocre” was acceptable; now “expensive but good” is the baseline requirement. If the actual experience doesn’t match the price, the backlash will be severe.

Also, developers are more price-sensitive than imagined. I’ve seen too many projects go from profitable to loss-making due to API costs, finally forced to switch models or cut features. An 83% increase may directly change technical choices for some small and medium developers.

But having said that, I think the biggest significance of Zhipu’s price increase isn’t the price itself, but the signal it sends to the industry:

Chinese LLMs can talk about value too, not just price-performance ratio.

This signal is important. If all domestic vendors are trapped in the mindset of “must be cheaper than OpenAI,” they can only be low-end alternatives forever. To truly participate in global competition, ultimately it depends on product strength.

As for whether GLM-5 is worth the price — I plan to test it myself. If it can indeed replace Claude Opus in actual projects, then this price is completely acceptable. If it’s just marketing hype, then Zhipu will fail this time.

What do you think? Is the price increase by Chinese LLMs now a sign of confidence, or are they digging their own graves?