GLM 5.3 API Pricing Drops 89% on Input Tokens — Z.ai Gets Competitive

What Changed

Z.ai has repriced the GLM 5.3 API. Input tokens dropped from $0.70 to $0.079 per 1M tokens — an -89% cut. Output tokens moved from $2.20 to $1.40 per 1M tokens, a meaningful but more modest reduction.

The asymmetry matters: if your workload is input-heavy (long context retrieval, document analysis, RAG pipelines with large prompts), the savings compound fast. Output-heavy tasks — chatbots, code generation, long-form drafting — see a real but less dramatic improvement.

Does It Matter for Your Stack?

At $0.079 input, GLM 5.3 moves into territory that invites direct comparison with budget-tier frontier models. For teams running high-volume inference — think millions of requests per day — the difference between $0.70 and $0.079 per 1M input tokens is not rounding error; it's a line item.

That said, price alone isn't the decision. GLM 5.3's benchmark profile, context window, and latency characteristics should be weighed against whatever you're currently running. Check the GLM 5.3 — live specs & price history page for up-to-date comparisons against alternatives indexed on this site.

Bottom Line

An 89% input price cut is a real signal — either Z.ai is chasing volume, responding to competitive pressure, or both. Either way, developers who dismissed GLM 5.3 on cost grounds have a reason to re-evaluate. Run your own token estimates against the new $0.079 input / $1.40 output rates before assuming a switch makes sense.

Ezra, Scout AI Team