GLM-5.2 becomes the fastest growing AI model of 2026 on a free MIT license

GLM-5.2 is the fastest growing AI model of 2026, and it is free to download.

Released by Chinese lab Z.ai on June 16, 2026 under a free MIT license, GLM-5.2 saw the fastest adoption of any model tracked by Vercel this year. In its first full week, daily token volume grew about 27x and its customer count grew about 80x.

Beating flagships at their own benchmark

Growth alone could be a curiosity. The benchmark result makes it a story. On SWE-bench Pro, a benchmark of real software engineering tasks, GLM-5.2 scores 62.1 percent, ahead of GPT-5.5 at 58.6 percent.

That means an open-weight model you can download and run yourself is now beating flagship US models at coding benchmarks, while costing a fraction of the price. The combination, not either fact alone, is what is pulling traffic so fast.

“Price is doing the work here”

Vercel’s head of agentic infrastructure, Harpreet Arora, describes the mechanics plainly: “Price is doing the work here.” When a task does not need the best model, teams route it to the cheapest one that is good enough.

That routing logic is quietly becoming the default architecture of production AI. Few teams pick one model for everything anymore. They classify tasks, send the hard ones to a frontier model, and push the volume to whatever clears the quality bar at the lowest cost. GLM-5.2’s numbers suggest it is clearing that bar for a lot of teams at once.

The “good enough” tier is going Chinese

The pattern extends beyond one model. The middle of the market, the workloads that need competence rather than brilliance, is increasingly served by open-weight Chinese models with permissive licenses and aggressive pricing. GLM-5.2’s MIT license removes even the friction of a vendor relationship: download it, run it, owe nobody anything.

For US labs, the challenge is sharpening. It is hard to charge premium prices for capability the open ecosystem gives away, and the open ecosystem keeps catching up faster. Have you shortlisted an open-weight model yet? The adoption curves say your competitors have.