Z.ai Raises GLM 5.2 API Prices by 41% — What It Costs You Now

The Numbers

Z.ai has repriced its GLM 5.2 API with a clean +41% increase across both token directions:

• Input: $0.545 → $0.76 per 1M tokens
• Output: $1.71 → $2.42 per 1M tokens

No phased rollout was announced — the new rates are live.

Does This Matter for Your Workload?

For low-volume hobby projects the delta is noise. For anything pushing serious token throughput, the math shifts meaningfully. A pipeline burning 100M input tokens and 50M output tokens per month goes from roughly $140 to $196 under the new pricing — a $56 monthly step-up before you've changed a single line of code.

Input-heavy workloads (bulk summarisation, document ingestion, RAG retrieval) feel the increase most, since the input rate moved proportionally the same as output but input tokens tend to dominate volume.

What to Do

If GLM 5.2 is a nice-to-have: now is a reasonable time to run a side-by-side against comparably-priced alternatives listed on this site. Check the GLM 5.2 — live specs & price history page for a current snapshot and competing models at similar capability tiers.

If GLM 5.2 is load-bearing: audit your input-to-output token ratio. High input ratios hurt more here. Consider caching repeated context aggressively — that unchanged $0.76 input rate compounds fast on redundant prompts.

If you're evaluating fresh: the +41% move makes the pre-increase assumption in any cost projection stale. Re-run your estimates with $0.76 input and $2.42 output before committing.

Price increases of this magnitude without a corresponding capability announcement are worth tracking — we'll flag any follow-up changes on the model page.

Ezra, Scout AI Team