GLM 5.2 API Pricing
Z.ai (GLM) · context window 1M · prices updated 2026-09-15
Where GLM 5.2 sits on price
At $4.40 per million output tokens, GLM 5.2 is a mid tier model — 2.1× the median output price of $2.08, which puts it cheaper than 34% and pricier than 66% of the models we track. It is the most expensive of the 17 Z.ai (GLM) models we track. Output costs 3.1× input, the usual spread — trim both, starting with the answer length.
Cached input is 90% cheaper. On an input-heavy workload you need roughly a 28% cache-hit rate to take 25% off the input line — reachable for chatbots and agents that resend the same system prompt and history.
What real workloads cost on GLM 5.2
These three workloads are the ones teams actually run on a mid-priced model — a $4.40/1M model is not bought for the same job as one ten times the price.
| Workload | Per request | Per month |
|---|---|---|
| Support chatbot — 500 conversations/day (5K in / 1.4K out) Conversation history re-sent each turn — the classic prompt-caching win. | $0.0132 | $197 |
| Document pipeline — 10K docs/day (3K in / 500 out) Batch-friendly: no user waiting, so a ~50% batch discount usually applies. | $0.0064 | $1920 |
| Long-context analysis — 200 runs/day (250K in / 4K out) Only possible on large-context models; input cost is nearly the whole bill. | $0.3676 | $2206 |
Model your exact traffic in the API cost calculator — it preloads GLM 5.2 with caching and batch options.
GLM 5.2 price history
Z.ai (GLM) last raised GLM 5.2's output price on Sep 15, 2026 (+105% per output token), from $0.6832/$2.15 to $1.40/$4.40 per 1M in/out. That is one of 18 repricings in the 44 days we have tracked it — worth knowing before you hard-code today's rate into a budget.
| Date | Input /1M | Output /1M | Output change |
|---|---|---|---|
| Aug 2, 2026 (tracking began) | $0.2842 | $0.8932 | — |
| Aug 3, 2026 | $0.6146 | $1.93 | +116% |
| Aug 4, 2026 | $0.76 | $2.42 | +25% |
| Aug 6, 2026 | $0.56 | $1.76 | -27% |
| Aug 7, 2026 | $0.6734 | $2.12 | +20% |
| Aug 8, 2026 | $0.252 | $0.792 | -63% |
| Aug 9, 2026 | $0.07 | $0.22 | -72% |
| Aug 10, 2026 | $0.76 | $2.42 | +1000% |
| Aug 12, 2026 | $0.5 | $3.15 | +30% |
| Aug 14, 2026 | $1.19 | $3.74 | +19% |
| Aug 15, 2026 | $0.49 | $1.54 | -59% |
| Aug 16, 2026 | $0.308 | $0.968 | -37% |
| Aug 17, 2026 | $1.19 | $3.74 | +286% |
| Aug 21, 2026 | $0.966 | $3.04 | -19% |
| Aug 25, 2026 | $1.19 | $3.74 | +23% |
| Sep 2, 2026 | $0.966 | $3.04 | -19% |
| Sep 12, 2026 | $0.6 | $2.00 | -34% |
| Sep 14, 2026 | $0.6832 | $2.15 | +7% |
| Sep 15, 2026 | $1.40 | $4.40 | +105% |
Change-points from our daily price snapshots (tracking since Aug 2, 2026; intraday moves between snapshots are not captured).
GLM 5.2 vs o4 Mini High
The closest-priced alternative from another vendor is o4 Mini High (OpenAI) — priced within a rounding error on output, with $0.3 less per million input tokens. When two models land this close on price, the decision is quality on your own workload, not the price sheet: run 50 real requests through both and compare.
See o4 Mini High pricing →Cheaper alternatives
More Z.ai (GLM) models
Frequently asked questions
› How much does the GLM 5.2 API cost?
GLM 5.2 costs $1.40 per million input tokens and $4.40 per million output tokens, with cached input at $0.14 per million (90% cheaper). That works out to roughly 227,273 output tokens per dollar.
› What does the support chatbot workload cost on GLM 5.2?
Support chatbot — 500 conversations/day (5K in / 1.4K out) costs about $0.0132 per request and $197 per month on GLM 5.2. Conversation history re-sent each turn — the classic prompt-caching win.
› What does it cost to fill GLM 5.2's 1M context window?
Sending 1M of input in a single request costs $1.47 at $1.40 per million tokens — before any output. With prompt caching that same fill drops to about $0.1468 on repeat requests. This is why large context windows are cheap to advertise and expensive to actually use.
› Is GLM 5.2 worth the price?
GLM 5.2 sits in the middle of the market (17 of 17 by price within Z.ai (GLM)). The honest test is a routing experiment: send the same 200 real requests to this model and to a tier below, and compare failure rate against the price gap — most teams find a majority of traffic never needed the pricier model.