GLM 5.2
GLM 5.2 processes million-token contexts for demanding text tasks.
Z.AILanguage ModelsClosedII 51.1
Function callingJSON modeStructured outputsReasoning
Model page Updated 2026-06-18
About GLM 5.2
GLM 5.2 is built as a closed-source LLM by Z.AI. Its architecture supports an exceptionally large context window of 1048576 tokens while remaining limited to text input and output. Parameter count details are not disclosed.
The model is suited for workloads that involve lengthy inputs such as full books, extended codebases, or multi-turn dialogues. Because it is not open-weight, access occurs through Z.AI's hosted API rather than local deployment.
Typical usage includes summarization of large documents, retrieval-augmented generation over long corpora, and complex reasoning chains that benefit from broad context retention.
Capabilities
Long-context reasoning
Code generation
Complex instruction following
Text summarization and analysis
Logical and multi-step reasoning
Multilingual text processing
Benchmarks & performance
Independent evaluation scores and measured speed.
51.1
Intelligence Index
50.7
Coding Index
75.9
Agentic Index
101
Tokens / sec
2.19s
Time to first token
Source: Artificial Analysis
How GLM 5.2 compares
GLM 5.2 (striped bar) vs other language models on intelligence, speed and price.
Intelligence
Artificial Analysis Intelligence Index · Higher is better · GLM 5.2 ranks #1 of 69
51
46
44
40
40
38
38
35
34
33
33
32
31
Speed
Output tokens per second · Higher is better · GLM 5.2 ranks #21 of 46
116
116
113
108
106
102
101
100
94
76
75
71
65
Price
USD per 1M output tokens · Lower is better · GLM 5.2 ranks #125 of 147
$2.5
$2.5
$3.0
Relace Search$3.0
$3.0
Llama 3.1 70B Hanami x1$3.1
$3.2
$3.3
$3.4
Switchpoint Router$3.8
$3.9
$3.9
$4.0
Sources: Artificial Analysis (intelligence, speed) · OpenRouter (price).
Best for
Long Document Analysis
Processes and reasons over full-length books, research papers, or extensive datasets within its 1M token context for accurate insights.
Enterprise Code Development
Generates, debugs, and refactors complex codebases while following detailed multi-step instructions across programming languages.
Global Content Localization
Performs multilingual text processing, summarization, and analysis for translating and adapting materials across languages with logical consistency.
Strengths & limitations
Strengths
- +Supports very large context windows
- +Strong performance on extended document tasks
- +Versatile across general text workloads
Limitations
- –Text-only modality
- –No vision or multimodal capabilities
- –Performance details beyond context length remain unverified
Pricing by provider
Live per-provider pricing & uptime, routed via OpenRouter. Prices are USD per 1M tokens.
| Provider | Input /1M | Output /1M | Context | Uptime |
|---|---|---|---|---|
| Wafer(fp4) | $1.20 | $3.20 | 203K | 78.9% |
| DeepInfra(fp4) | $1.20 | $4.20 | 1049K | 75.4% |
| Phala | $1.40 | $4.40 | 524K | 91.8% |
| Cloudflare | $1.40 | $4.40 | 262K | 99.6% |
| Fireworks | $1.40 | $4.40 | 1049K | 99.3% |
| Z.AI(fp8) | $1.40 | $4.40 | 1049K | 99.7% |
| Friendli | $1.40 | $4.40 | 1049K | 99.8% |
| Parasail(fp8) | $1.40 | $4.40 | 1049K | 88.9% |
| Novita(fp8) | $1.40 | $4.40 | 1049K | 99.8% |
| AtlasCloud(fp8) | $1.40 | $4.40 | 203K | 96.6% |
| StreamLake | $1.40 | $4.40 | 1024K | 96.9% |
| Io Net(fp8) | $1.68 | $5.28 | 262K | 90.1% |
Cost calculator
Estimate what GLM 5.2 would cost for your usage.
$0.00280
per request
$28
estimated / month
Based on GLM 5.2's $1.20/1M input · $3.20/1M output. Estimate only — actual cost varies by provider and caching.
Quick start
OpenRouter's API is OpenAI-compatible — most SDKs work by just swapping the base URL. Only the model slug changes between models.
JavaScript · openai
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://openrouter.ai/api/v1",
apiKey: process.env.OPENROUTER_API_KEY,
});
const completion = await client.chat.completions.create({
model: "z-ai/glm-5.2",
messages: [{ role: "user", content: "Hello!" }],
});
console.log(completion.choices[0].message.content);Model slug: z-ai/glm-5.2
Editor's verdict
Our take on GLM 5.2
GLM 5.2 is Z.AI's proprietary language models with a 1049K-token context window.
On independent testing it scores 51.1 on the Artificial Analysis Intelligence Index, running at roughly 101 tokens per second with about 2.19s to first token.
At $3.20 per 1M output tokens, it is mid-priced for its class, served by 12 providers.
It is available through Z.AI's API and aggregators like OpenRouter.
Best suited to supports very large context windows and strong performance on extended document tasks.
Did you find this helpful?
Frequently asked questions
GLM 5.2 provides a context window of 1048576 tokens.
User reviews
Real, verified reviews from the community shape this model's rating.
Loading reviews…
Other GLM models
Sibling versions in the GLM family from Z.AI.
Similar models
Other language models worth comparing.