Which Zhipu AI model is the cheapest?

GLM-4.5 Flash is the most affordable general text model, priced at $0.019 per 1 million input tokens and $0.078 per 1 million output tokens via GPUniq.

How can I get a 30% discount on Zhipu API calls?

You can receive a 30% discount on specific models like GLM-4.7 Flash, GLM-4.5 Air, and GLM-4.6V by routing your API requests through the GPUniq gateway.

What is the difference between standard and reasoning model pricing?

Standard models like GLM-4.5 Flash are optimized for speed and low cost, while reasoning models like GLM-4.6 Thinking are priced higher—at $0.780 input and $1.560 output—to account for complex compute requirements.

Does Zhipu AI charge differently for input and output tokens?

Yes, Zhipu AI bills for input and output tokens separately, with output tokens generally costing more across most models to cover the compute resources required for generation.

Zhipu AI API Pricing Guide and Cost Reduction Tips

Zhipu AI API Pricing: Cut Costs by 30%. A macro close-up of a single high-end GPU chip illuminated by a sharp, cold blue light, resting against a backdrop of neatly organized, tangled fiber-optic cables that transition from a chaotic mess — GPUniq

Data as of October 6, 20261 source

Drafted with AI tools and edited by a human. The figures and conclusions were checked by a GPUniq editor.

TL;DR

GLM models cost between $0.019 and $1.336 per 1M input tokens via GPUniq infrastructure. Discounts reach up to 30% off standard vendor list prices. GLM-4.5 Flash is the lowest-cost text model at $0.019 input and $0.078 output per 1M tokens.

What Does Zhipu AI API Access Cost?

Zhipu AI bills API requests on token volume. You pay separately for input prompt tokens and generated output tokens. Output tokens cost more on almost all models to cover generation compute.

Routing API calls through GPUniq drops costs below standard vendor rates. Baseline GLM models get a 7.5% price reduction. Selected endpoints get a 30% discount.

{
  "model": "glm-4.5-flash",
  "messages": [
    {"role": "system", "content": "You are a concise assistant."},
    {"role": "user", "content": "Summarize standard API token pricing metrics."}
  ]
}

Model rates depend on context capacity, processing speed, and architectural target. Fast text endpoints cost just cents per million tokens. Reasoning endpoints push higher.

Which Zhipu Model Is the Cheapest?

GLM-4.5 Flash is the lowest-cost general text model in Zhipu's catalog. It bills at $0.019 per 1M input tokens (1.91 RUB) and $0.078 per 1M output tokens (7.65 RUB). That rate includes a 7.5% discount via GPUniq.

GLM-4.5 Flash Token Cost:
Input:  $0.019 / 1M tokens (1.91 RUB)
Output: $0.078 / 1M tokens (7.65 RUB)

GLM-4.5 Flash retains a 128k context window. Use it for high-volume pipelines where speed and unit economics come first.

Common deployment targets:

Basic routing and intent classification.
Text categorization.
Parsing structured plain text documents.
Initial chat turn handling.

GLM OCR charges $0.054 per 1M tokens (5.26 RUB) flat for both input and output across a 128k context window.

How Do GLM Model Costs Compare?

Zhipu structures endpoints into fast, balanced, and reasoning tiers. Most standard models use 128k context windows. GLM-4.5 Air and GLM-4.6V expand to 131k tokens. GLM-4.7 Flash expands to 200k tokens.

GLM-4 Air costs $0.068 per 1M input tokens and $0.068 per 1M output tokens (6.70 RUB). GLM-4.5 Air charges $0.195 input and $1.275 output per 1M tokens, but carries a 30% discount.

Model	Category	Context (k)	Input ($/1M)	Output ($/1M)	Discount vs List
GLM-4.5 Flash	fast	128	$0.019	$0.078	7.5%
GLM-3 Turbo	fast	128	$0.068	$0.068	7.5%
GLM-4 Air	fast	128	$0.068	$0.068	7.5%
GLM-4.7 Flash	fast	200	$0.090	$0.600	30.0%
GLM-4 AirX	balanced	128	$0.137	$0.546	7.5%
GLM-4 Long	balanced	128	$0.137	$0.546	7.5%
GLM-4.5 Air	balanced	131	$0.195	$1.275	30.0%
GLM-4.6V	balanced	131	$0.450	$1.350	30.0%
GLM-4	balanced	128	$0.556	$1.111	7.5%
GLM-4.6 Thinking	reasoning	128	$0.780	$1.560	7.5%

Reasoning endpoints sit at the top of the price scale. GLM-4.6 Thinking costs $0.780 input (76.53 RUB) and $1.560 output (153.07 RUB) per 1M tokens. GLM-4.1V Thinking Flash offers a reasoning-tier option at $0.125 input and $0.714 output.

Where Can You Get a 30% Discount?

GPUniq applies tier discounts directly at the routing layer. Standard endpoints run at a 7.5% reduction. Three modern GLM models hit the full 30% discount:

Model	Category	Context (k)	Input ($/1M)	Output ($/1M)
GLM-4.7 Flash	fast	200	$0.090	$0.600
GLM-4.5 Air	balanced	131	$0.195	$1.275
GLM-4.6V	balanced	131	$0.450	$1.350

Endpoints taking the 7.5% discount include GLM-4.5 Flash, GLM-4 Air, GLM-3 Turbo, GLM-4, GLM-4 AirX, GLM-4 Long, GLM-4.6 Thinking, and GLM-4V Plus.

Price Calculation Example (GLM-4.7 Flash Input):
Discounted Rate (30% off list): $0.090 / 1M tokens (8.83 RUB)
Standard List Equivalent:       $0.1286 / 1M tokens

No commit minimums apply. Unified endpoints assign the discount automatically.

How Does Vision API Pricing Work?

GLM-4.6V and GLM-4V Plus are offered in the balanced catalog tier. GLM-4.6V provides 131k context at $0.450 per 1M input tokens (44.15 RUB) and $1.350 per 1M output tokens (132.46 RUB). It receives a 30% discount off list rates.

Zhipu Endpoint Pricing (GPUniq Rates):
GLM-4.6V (30% Discount):     $0.450 Input / $1.350 Output per 1M
GLM-4V Plus (7.5% Discount): $1.336 Input / $1.336 Output per 1M

GLM-4V Plus features a 128k context window and bills at $1.336 per 1M input tokens and $1.336 per 1M output tokens (131.05 RUB). GLM-4.6V cuts input costs by 66.3% compared to GLM-4V Plus.

How Can You Reduce Zhipu API Expenses?

Combining smart routing with gateway discounts lowers token spend.

+-------------------------------------------------------------+
|                      Incoming Prompt                        |
+-------------------------------------------------------------+
                               |
                               v
               /-------------------------------\
              / Is the task simple or complex?  \
             \---------------------------------/
                   /                       \
          (Simple) /                         \ (Complex/Reasoning)
                  v                           v
     +-------------------------+     +-------------------------+
     |     GLM-4.5 Flash       |     |   GLM-4.5 Air / 4.6   |
     | $0.019 In / $0.078 Out  |     | 30% Discount Gateway    |
     +-------------------------+     +-------------------------+

Optimizing spend across pipelines:

Route simple steps to GLM-4.5 Flash ($0.019 input / $0.078 output). Save GLM-4.6 Thinking ($0.780 input / $1.560 output) for multi-step reasoning.
Trim fixed system instructions. Input tokens bill on every API hit.
Use GLM-4.7 Flash ($0.090 input / $0.600 output) for long context runs up to 200k tokens. It gets the 30% discount rate.
Point SDK calls to GPUniq gateway endpoints to lock in discounts between 7.5% and 30%.

Methodology

Prices are GPUniq's public retail rates per million tokens, read from the live model catalog at the moment of the snapshot. Ruble figures use the rate GPUniq applies to ruble top-ups at that moment. The discount is against the model vendor's own list price.